Applications of comprehensive metabolic profiling, broadly termed “metabolomics”, are increasingly common in epidemiology and genetics. This is due to recent developments in quantitative methodologies and various appealing results from their applications on understanding life-course health and disease etiologies. Most epidemiological studies routinely analyse blood biomarkers, for example total cholesterol and glucose. Our team has developed a high-throughput serum NMR metabolomics platform that measures the concentrations of multiple standard biomarkers but in the same time also provides quantitative molecular data on over 200 metabolic measures not commonly studied in epidemiology. So far we have used the platform to analyze close to 200,000 samples and the molecular data have been used in various epidemiological and genetics studies. Some of the research areas are briefly described here. Evidently, the increasing amounts of quantitative data on systemic metabolism are revealing a plethora of novel molecular biomarkers suggesting that quantitative metabolomics will eventually reshape the way in which epidemiology and genetics are practiced in the near future.
Our team has developed an automated high-throughput serum NMR metabolomics platform (click here for the detailed description of the platform) that provides quantitative molecular data on 14 lipoprotein subclasses, their lipid concentrations and composition, various fatty acids, α-acid glycoprotein as well as on numerous low-molecular-weight metabolites, including amino acids, glycolysis related measures and ketone bodies (click here for the detailed information on the molecular measures). The molar concentrations of these measures are obtained from a single serum sample with costs comparable to standard lipid measurements. These molecular data have been used to study type 1 and type 2 diabetes aetiology as well as to characterize the molecular reflections of the metabolic syndrome, long-term physical activity, diet and lipoprotein metabolism. The results have revealed new biomarkers for early atherosclerosis, type 2 diabetes, diabetic nephropathy, cardiovascular disease and all-cause mortality.
The quantitative molecular data are obtained from a single serum or plasma sample via 3 molecular windows, each of which is optimised on the basis of current automated NMR technology to provide comprehensive quantitative data on particular molecular entities; two windows are used to acquire data on lipoprotein subclasses and low-molecular-weight metabolites in native samples. After these measurements the lipids in the sample are extracted and a third window analysed for the information on individual lipids and fatty acids. The platform is highly automated with respect to experimentation as well as spectral data analysis and provides quantitative data on the molar concentrations of over 200 metabolic measures (click here for the detailed information on the molecular measures). Hundreds of biologically relevant derived measures (e.g., metabolite ratios as proxies for enzymatic activity) can also be calculated.
The cost of this analysis is comparable to standard lipid measurements. We have profiled blood specimens from close to 200,000 individuals from large cohorts, e.g., in Finland, Estonia and UK (click here for the key collaborators).
The accumulating new information is enhancing our knowledge on the systemic manifestations of the gradual transition between health and disease. A particular emphasis is on metabolic signatures of risk factors among healthy people and the metabolic changes during the life course in order to be able to pinpoint subclinical disease. The improved understanding of the molecular mechanisms and pathways leading to cardiometabolic diseases will benefit translational medicine and aid personalization of dietary or medical interventions.
The genetic makeup of an individual is influencing the comprehensive metabolite profile. By analysing the genetic determinants of the individual metabolites in large population studies, we have discovered novel genes involved in regulating the systemic metabolism. Common genetic variants were identified for the concentrations of blood lipids, fatty acids and amino acids. This information improves our understanding of the role of the metabolites. Twin studies further indicate substantial heritability of the metabolic profile. This illustrates that we are only at the beginning to uncover the genetic basis for the variations in systemic metabolism.
The risks for the development of metabolic diseases are partly due to the individual genetic setup. Many genes have been linked with metabolic diseases in genome-wide association studies; however, the mechanisms underlying how the genes predispose to disease are often unclear. Combining comprehensive quantitative metabolite profiling with population genetics gives insights into the functional roles of genes involved in various physiological as well as disease processes. The individual molecular measures can be used as intermediates connecting the genes with the disease phenotypes. We have studied associations between the metabolite profile and genes firmly linked, e.g., with lipid metabolism, diabetes, hypertension and liver dysfunction. Our results contribute to enhanced molecular understanding of novel genes linked with cardiometabolic diseases.
Quantitative metabolomics has identified biomarkers associated with metabolic risk factors and disease risk. However, addressing the causal relation between the metabolites, risk factors, and disease outcomes is a key challenge in epidemiology. Causality can be inferred based on the concept of Mendelian randomisation, where genetic variants are used as proxies for the metabolites. Effective application of Mendelian randomisation typically calls for a very large study size; the accumulating quantitative data on systemic metabolism will therefore provide a unique and powerful resource for future studies to address the issues of causality. In addition, we have also recently illustrated that the comprehensive metabolic profile can effectively be used to address genetic pleiotropy and thereby assist in the selection of good genetic instruments.
The development of atherosclerosis, the culprit of cardiovascular diseases, begins early in life, typically several decades before clinical manifestations. Improved prediction of cardiovascular risk in early adulthood would enable refining primary prevention and thereby assist in individual health maintenance as well as in saving health care costs.
Via comprehensive metabolite profiling we have identified novel systemic biomarkers for the early progression of atherosclerosis in young adults. These results provide novel insights into the molecular mechanisms underlying the pathogenesis of atherosclerosis and improve the early assessment of cardiovascular disease risk. Our findings also indicate that quantitative serum NMR metabolomics has potential to benefit individualised treatment strategies and the prevention of cardiovascular events in a very cost-effective manner.
We have extended the studies of systemic biomarkers to later stages of cardiovascular disease in large population cohorts. We have indicated that serum concentrations of polyunsaturated fatty acids, including docosahexaenoic acid, are independent predictors of subclinical atherosclerosis and cardiovascular risk. Our findings denote that aromatic amino acids, glycolysis substrates and polyunsaturated fatty acids are reflective of cardiovascular disease risk in general population settings. Extensive work in this area is on-going and it is too early to conclude if these biomarkers will be helpful in refining risk predictions with a single prediction model for the whole population. However, we already know that these biomarkers highlighted by quantitative metabolomics appear consistent in multiple populations and as assayed by NMR and MS profiling technologies.
Proton NMR spectroscopy enables quantification of lipoprotein subclasses in a cost-effective manner. Our methodology provides extensive information on 14 lipoprotein subclasses, their particle concentrations, lipids and lipid composition. These are a treasure trove of enhanced understanding of lipid metabolism in health and disease. We have, for example, used this approach to characterise the lipid metabolism of patients with type 1 and type 2 diabetes as well as to provide metabolic fine mapping of genes implicated in lipid metabolism. We have also examined the lipoprotein subclass profiles with various metabolic risk factors such as obesity and the metabolic syndrome as well as studied dietary effects.
Insulin resistance is a core defect in the development of type 2 diabetes. Already early in life, insulin resistance is associated with an adverse lipid profile and increased risk for future diabetes. However, insulin resistance is leading to much wider alterations of the metabolite profile than the conventional characteristics of the metabolic syndrome would indicate. We have examined metabolic signatures of insulin resistance in several large population studies. The lipoprotein subclass profile, the fatty acid composition, ketone bodies as well as amino acid concentrations measured from a fasting blood sample are strong indicators of the current degree of insulin resistance. Some metabolites are also reflective of the risk for impaired insulin sensitivity later in life. The manifestations of insulin resistance in the systemic metabolite profile are seen already in adolescents and young adults free of impaired glucose tolerance.
Type 2 diabetes affects hundreds of millions people worldwide. The disease accounts for around 10% of the total health care costs in Finland. Diabetes leads to many complications in the eyes, kidneys, arteries and heart. The diagnosis of diabetes is preceded by elevated glucose levels both in the fasting and postprandial state. Metabolomics profiling has demonstrated that the continuum of prediabetes and diabetes is widely reflected in the comprehensive metabolite profile. Many of the metabolites quantified by high-throughput profiling are predictors of future development of hyperglycaemia and overt diabetes.
Type 1 diabetes follows the autoimmune destruction of pancreatic beta-cells at young age. In addition to perturbations in the glycaemic profile, type 1 diabetes also affects lipoprotein and lipid metabolism. Diabetic nephropathy is the most serious complication of type 1 diabetes. There is no treatment to protect the kidneys from poorly controlled diabetes and thereby prevention of the initial metabolic insults is currently the only effective approach to reduce the high mortality related to diabetic nephropathy.
We have applied quantitative serum metabolomics to study systemic metabolism in a large cohort of Finnish type 1 diabetes patients. We have applied metabolomics both in the examination of single discriminatory biomarkers (such as sphingomyelin as a predictor of kidney disease) as well as for the data-driven characterisation of high-risk metabolite patterns (such as the lipoprotein subclass profile). Our results illustrate that quantitative metabolic phenotyping brings us one step closer to appreciate the unique set of regulatory perturbations that predispose to kidney injury and paves the way for multiparametric risk assessment in type 1 diabetes.
Physical activity reduces the risk of diabetes and coronary heart disease. Investigating the biology of physical activity is challenging since the effects are modulated through multiple interconnected pathways, which are dependent on both genetics and lifestyle. Using the quantitative metabolomics approach and a combination of twin and population-based studies, we have demonstrated associations of a plethora of systemic metabolic measures with long-term physical activity. Our results demonstrate consistent effects across multiple metabolic pathways, with all the systemic metabolic consequences of physical activity being positive from the cardiovascular health point of view.
Diet is a key regulator of systemic metabolism. Quantitative serum metabolomics provides an exemplary methodology to concomitantly study multiple molecular pathways reflective of diet or dietary interventions. We have illustrated associations of protein diet with circulating amino acids profiles and the composition of fatty acids. Interestingly, we noted in a berry intervention study that the overall effects of berries on serum metabolome are affected by the individuals’ baseline metabolome. This finding opens interesting and novel avenues for clinical dietary interventions, not to forget drug treatments.
Professor Ala-Korpela is one of the pioneers in applying NMR spectroscopy in biomedicine to study lipoproteins and body fluids. He published an authoritative review on 1H NMR spectroscopy of human blood plasma in Progress in Nuclear Magnetic Resonance Spectroscopy already in 1995 (Vol 27; pages 475-554). In addition, he has co-authored reviews in the area of lipoprotein modifications and early atherosclerosis mechanisms. An extensive work, entitled "Structure of low density lipoprotein (LDL) particles: basis for understanding molecular changes in modified LDL." was published in 2000 in Biochimica Biophysica Acta - Molecular and Cell Biology of Lipids (Vol 1488 pages 189-210); this paper presented novel molecular representations of lipoprotein particles and has been continuingly cited since then. The presented molecular models of lipoprotein particles were updated in 2008 to refine the hydrophobic lipid distributions in the particles (Chemistry and Physics of Lipids, Vol 155; pages 57-62).
Absolute quantification of NMR data is crucial for the modern metabolomics applications. Prof Ala-Korpela’s early scientific work in the 90’s focused on developing and applying sophisticated model line shape fitting in biomedical NMR spectroscopy. At 2001 he published with Dr Mierisová an authoritative review entitled "MR spectroscopy quantitation: a review of frequency domain methods" in NMR in Biomedicine (Vol 14; pages 247-59).
Around mid 2000 Prof Ala-Korpela started to focus more on metabolic and clinical risk assessment and published a review in 2006, together with Dr Sipola and Prof Kaski, on “Characterization and molecular detection of atherothrombosis by magnetic resonance – potential tools for individual risk assessment and diagnostics” in Annals of Medicine (Vol 38; pages 322-36). This review was one the first from Prof Ala-Korpela’s new Computational Medicine Team and has then been followed by his authoritative, invited reviews “Potential role of body fluid 1H NMR metabonomics as a prognostic and diagnostic tool” in 2007 in Expert Review of Molecular Diagnostics (Vol 7; pages 761-73) and “Critical evaluation of 1H NMR metabonomics of serum as a methodology for disease risk assessment and diagnostics” in 2008 in Clinical Chemistry and Laboratory Medicine (Vol 46; pages 27-42).
In 2011 Prof Ala-Korpela was a lead guest editor for a special issue on “Clinical and epidemiological metabonomics” in the Journal of Biomedicine and Biotechnology. Also in 2011, a commissioned opinion entitled “Genome-wide association studies and systems biology: Together at last”, was published in Trends in Genetics (Vol 27; 493-8) in collaboration with Mr Kangas and Dr Inouye. A musing, entitled “Quantitative high-throughput metabolomics: a new era in epidemiology and genetics” was commissioned in 2012 for a special issue on disease metabolomics in Genome Medicine (Vol 4; nro 36), contributed also by Mr Kangas and Dr Soininen, long-term key members of the Computational Medicine Research Team.
An authoritative invited review on “Quantitative serum nuclear magnetic resonance metabolomics in cardiovascular epidemiology and genetics” is expected to appear soon.
Some key punch lines from the recent reviews and opinions include:
We envision that quantitative high-throughput NMR metabolomics will be incorporated as a routine in large biobanks; this would make perfect sense both from the biological research and cost point of view – the standard output of over 200 molecular measures would vastly extend the relevance of the sample collections and make many separate clinical chemistry assays redundant.
(2014, under review)
The increasing amounts of quantitative data on systemic metabolism are revealing a plethora of novel molecular biomarkers suggesting that quantitative metabolomics will eventually reshape the way how epidemiology and genetics are practiced in the near future.
(2014, under review)
...NMR is currently the only methodology capable of offering reproducible high-throughput metabolite quantifications in a cost-effective manner.
(2014, under review)
It is of particular note that sound study design in epidemiological metabolomics follows similar principles as standard epidemiology. Metabolomics technologies can also be directly applied in already existing large epidemiological sample collections. There is no need for a particular study design to apply quantitative metabolomics. Notably, quantitative metabolomics data is in fact identical to an extensive collection of molecular concentrations.
(2014, under review)
Quantitative metabolomics allows both testing clear hypotheses and hypotheses-free assessment of biological questions.
(2014, under review)
We are phasing an era when extensive sets of samples can be profiled by NMR metabolomics in a short time scale.
(2014, under review)
...NMR-based metabolomics will enable longer, healthier and happier lives of people.
(2014, under review)
...we should inherently accept the biologically inevitable metabolic and disease continuum instead of being hampered by the apparently unattainable black-and-white diagnostics. Common disorders, being multigenic, are essentially quantitative traits, and ultimately we cannot hide from this fundamental feature of nature.
(2012, Genome Medicine )
...metabolomics will be truly useful in epidemiology or in genetic studies only if quantitative data on specific, identified metabolites are available. (2012, Genome Medicine)
We expect to see greater genetic homogeneity when composite clinical measures are separated into distinct biochemical components that more closely reflect underlying molecular pathways. (2011, Trends in Genetics);
A key strategic reason to use proton NMR spectroscopy to analyse lipoprotein subclasses is the avoidance of their tedious physical isolation from plasma and the consequent potential for detailed studies of extensive populations. (2008, CCLM)
It is essential to make a difference between a diagnosis and an individual risk assessment, particularly when dealing with diseases, such as atherosclerosis in which the borderline between health and disease is intrinsically indistinct. (2008, CCLM)
Aetiology of atherothrombosis – natural fuzziness between health and disease. (2008, CCLM)
A practical approach to personalized medicine via proton NMR metabonomics would be the biological (sub)classification of individuals with respect to their (global and local) metabolic states, for example, to assess the risk for developing CHD or diabetes, as well as with reference to various future options, such as continuing the current lifestyle and nutrition, tailored lifestyle changes or prescription of particular medications. (2007, Expert Rev Mol Diagn)
The use of in vitro proton magnetic resonance spectroscopy (MRS), to detect lipoprotein subclasses and plasma metabolites, together with in vivo multi-contrast MRI, to characterize plaque components, could provide a genuinely new approach to uncover individual intermediate atherothrombosis before any clinical manifestations. (2006, Ann Med)
...maybe in the near future we will see examination requests, for instance, of the form: complete plasma lipid and metabolite analysis by proton NMR. (1995, PNMRS)