Research Group at Laboratory of Computational Engineering


1H NMR Metabonomics for Disease Risk Assessment

Researchers: Jukka Heikkonen, Johanna Hokkanen, Antti Kangas, Kimmo Kaski, Linda Kumpula, Niko Lankinen, Ville-Petteri Mäkinen, Jaakko Niemi, Aino Salminen, Jari Saramäki, Teemu Suna, Taru Tukiainen, Pasi Soininen (a), Reino Laatikainen (a), Matti Jauhiainen (b), Petri Ingman (c), Sanna Mäkelä (d), Antti Nissinen (d), Minna Hannuksela (d), Markku Savolainen (d), Marja-Riitta Taskinen (e), Per-Henrik Groop (f), Timo Liimatainen (g), Petri Sipola (h), Mika Ala-Korpela*
(a) Department of Chemistry, University of Kuopio;
(b) Department of Molecular Medicine, National Public Health Institute, Helsinki;
(c) Department of Chemistry, Instrument Centre, University of Turku;
(d) Department of Internal Medicine, University of Oulu;
(e) Division of Cardiology, Department of Medicine, University of Helsinki;
(f) Folkhälsan Institute of Genetics, Folkhälsan Research Centre, University of Helsinki;
(g) A. I. Virtanen Institute, Biomedical NMR Research Group, University of Kuopio,
(h) Department of Clinical Radiology, Kuopio University Hospital.)
- *Correspondence to

The genomics, transcriptomics and proteomics represent the genome-oriented main discipline in life sciences. Physiology constitutes the triggering of specific functional pathways by environmental signals and thereby, the phenotype of a biological system is largely reflected by its metabolite composition and their interactions. An essential and complementary ‘omics’-approach in understanding of biomolecular function is therefore metabonomics – the quantitative measurement of the time-related multiparametric metabolic responses of multicellular systems to pathophysiological stimuli or genetic modification.

Measuring metabolites is not new. For decades, clinicians have charted chemistries in blood, urine, and other body fluids — e.g., using glucose to track diabetes and cholesterol to monitor heart disease. What is new in the metabonomics approach is that we are now casting a wider net, attempting to gather an unbiased sample of metabolites that can serve as a snapshot of an organism's physiology. We could also talk about ‘global biochemistry’. The ultimate goal of metabonomics is to be able to distinguish between an individual who is healthy and someone who has (diagnosis) — or might develop (risk assessment) — a disease. In the field of metabonomics, mass spectrometry and NMR spectroscopy have become the two key technologies. An appealing feature of NMR spectroscopy for metabonomic applications is its specific yet non-selective nature (see Figure 1). Particularly, 1H NMR has the advantage of efficiently obtaining information on large numbers of metabolites in biofluids in vitro as well as in various tissues ex vivo and in vivo.

Figure 1
Figure 1: Illustration of characteristic 1H NMR molecular windows for a type 1 diabetic (T1DM) patient. The assignments for the LIPO window resonances refer to fatty acids in triglycerides, cholesterol compounds and phospholipids in various lipoprotein particles, the cholesterol backbone –C(18)H3 and the –N(CH3)3 groups of surface phospholipids. The LMWM resonances marked gp are from the N-acetyl protons of mobile N-acetylated carbohydrate side-chains of glycoproteins.

Our biomedical focus is understanding atherothrombosis – the various complex life long processes of harmful lipoprotein particle elevation and their modifications leading to lipid accumulation and potentially to a thrombus formation and a subsequent heart attack. Applications of 1H NMR metabonomics to study human serum are experimentally rather fast and straightforward. Measuring lipoprotein subclass profiles by 1H NMR therefore contrasts favourably to other lipoprotein measurement protocols and is currently receiving wide academic and commercial interest. The independent role of lipoprotein subclasses for the risk assessment and development of atherothrombosis is currently well recognised. We have indeed recently illustrated the inherent suitability of 1H NMR metabonomics for automated studies of lipoprotein subclass related metabolic interactions in a clinically relevant context and demonstrated the power of self-organising map (SOM) analysis in an extensive and representative case of 1H NMR metabonomics (see Figure 2).

Figure 2

Figure 2: A schematic simplification of the challenge related to the risk assessment and diagnosis of atherothrombosis.

Atherosclerosis is a diffuse systemic disease that is characterised by the local build-up of lipid-rich plaques within the walls of large arteries. The atherothrombotic processes are multigenetic, being influenced also by dietary and environmental components, and are apparent as early as the second decade in life with an increased incidence in the elderly. Atherothrombosis involves inflammatory processes with an array of metabolic, molecular and cellular manifestations in tissues, e.g., those depicted within the arterial wall in Figure 2. A varying degree of these intimal processes are reflected by the biochemistry of body fluids, such as serum. The biological heterogeneity as well as the slow development and progression of pathological conditions make the borderline between ‘health’ and ‘disease’ indistinct. One option to approach the problem, as previously presented by us, is 1H NMR metabonomics of serum equipped with a chemometric classifier, e.g., a SOM. On the left in Figure 2 a hypothetical SOM is shown together with four overlapping clusters that are thought to represent the metabolic changes in the arterial intima. While definite classification as ‘healthy’ and ‘diseased’ may not be available by nature, the metabonomics approach with a holistic look at the multidimensional metabolic changes may prove useful in the assessment and follow up of an individual ‘health path’ (represented by the light green line within the SOM) alongside the interplay between metabolic pathways and their consequences.

It is our main aim to perform extensive metabonomic NMR studies in various clinically relevant sets of serum samples and to develop as well as to apply data analysis approaches capable of detecting differences in the biomolecular status of the individuals in relation to disease risk assessment and diagnosis. We have started our clinically oriented NMR metabonomic applications within two projects: i) alcohol related anti-atherogenic processes and diseases and type 1 diabetes and the risk of diabetic nephropathy and vascular complications. The studies involve systematic linking of various biochemical and biophysical methodologies to study atherothrombosis and lipoprotein related phenomena. This coalition integrates currently approximately 30 scientists, from 10 different institutions, working in close collaboration. Within these studies we will put particular effort towards a comprehensive systems biology approach in which we will be integrating complementary metabolic data from NMR spectroscopy with all available other data of clinical significance (including various clinically utilised biochemical markers, diagnostic data, other spectroscopic data and genetic information). The general idea is illustrated in Figure 3.

Figure 3

Figure 3: Illustration of the statistically significant (P < 0.001) Spearman correlations between the biochemical variables and the resonances in the 1H NMR spectra of serum in type 1 diabetic patients in the LIPO (top) and LMWM (bottom) windows within the frequency regions of 0.4 – 3.3 ppm. A colour bar for the correlation coefficients is shown on the top.

Our first results demonstrate that 1H NMR metabonomics clearly distinguishes metabolic characteristics of T1DM and appears approximately as good means to diagnose diabetic nephropathy from serum as an advanced set of biochemical variables. A presentation by Ph.D. student Ville-Petteri Mäkinen, related to this topic, won the 2nd prise in the Young Investigator Award competition at the 23rd Annual Scientific Meeting of the European Society for Magnetic Resonance in Medicine and Biology (ESMRMB), September 21–23, Warsaw, Poland, 2006.

Along the 1H NMR experimentation, we have also focused on developing biochemistry based signal models for the key metabolites in human serum and plasma and setting up a software allowing to simulate sets of 1H NMR spectra corresponding to various biomedical conditions, such as different lipoprotein subclass profiles and metabolic pathways, related to the risk assessment of coronary heart disease. The inherent capability of 1H NMR to quantify lipoprotein subclasses was found to depend on the particular subclass. Ten-fold differences in the quantification accuracy were observed between the most accurate subclasses (VLDL and HDL3) and the least accurate subclass (LDL1). These finding are noteworthy in relation to the clinical utility of lipoprotein subclass quantification by 1H NMR spectroscopy. We have also designed a potential scheme for the utilisation of MR methodologies in the assessment of long-term and short-term risk for atherothrombotic events. This scheme can be seen as one option to elucidate the potential of MR in detecting individual intermediate atherothrombotic end points and utilising their prognostic value before the occurrence of a definite end point; see Figure 4.

In collaboration with the Biomedical NMR Research Group at the University of Kuopio, we have presented a novel identification and analysis of mobile cholesterol compounds in an experimental glioma model by 1H MRS in vivo. The introduced 1H MRS approach facilitates a non-invasive follow-up of mobile cholesterol compounds, paving way for studies of tumour cholesterol metabolism in vivo. The inclusion of the multi-component model lineshape of cholesterol backbone into spectral analyses leads to novel biochemical information on cholesterol compounds in vivo, and provides a more reliable analysis of the overlapping fatty acid resonances.

Our recent results have shown that the new 1H NMR metabonomics approach, combined with newly developed multi-factorial statistical analysis methods, is capable of finding clear metabolite alterations in the serum samples of various patient groups. The spectra contain an unforeseen wealth of information on serum metabolites, for example, in relation to type 1 diabetes and diabetic nephropathy. These data provide a systems biology view on metabolism, an aspect that cannot be captured by a handful of conventional clinical variables. It is our intention to develop NMR metabonomics towards increased biomolecular understanding and thereby to improve individual disease risk assessment and to develop more specific molecular markers for the detection and follow-up of atherothrombosis.

The work in the near future will focus on collecting more clinically relevant data via 1H NMR metabonomics of serum and on related spectral analyses. We will study the role of 1H NMR spectroscopy in lipoprotein subclass quantification and the performance of various data analysis methods in biomedical spectroscopy applications.

Figure 4

Figure 4: A potential scheme utilising MR methodologies in the risk assessment of long-term risk for atherothrombotic events (non-symptomatic individuals) and of short-term risk for recurrent cardiovascular events after an experienced acute coronary syndrome (ACS) (symptomatic patients). At risk assessment point I the molecular constituents of serum, including lipoprotein subclasses, could be assessed by in vitro 1H MRS metabonomics for non-symptomatic individuals. If high long-term risk for atherothrombotic events is indicated, non-invasive in vivo MRI could follow for the potential detection of plaque (risk assessment point IIa) and subsequent compositional evaluation of the vulnerability of the detected plaque(s) for rupture or erosion (IIb). Depending on the outcome from the plaque detection and assessment by MRI the individual could accordingly be directed for further actions. If vulnerable plaque at point IIb would be detected, considerations for aggressive drug therapy or invasive therapies such as angiographic stenting or bypass surgery would be needed. In the case of an individual with an experienced ACS (III) in vitro 1H MRS metabonomics could be used to complement the clinical protocols when evaluating the risk for recurrent cardiovascular events and the proper individual treatment options. For some symptomatic patients in vivo MRI might also be feasible at point III for direct assessment of plaque composition and vulnerability. This scheme can be seen as one option to elucidate the potential of MR in detecting individual intermediate atherothrombotic end points and utilising their prognostic value before the occurrence of a definite end point. The recent MR findings and developments awaken confidence that this kind of schemes might be operational in the near future saving both human suffering and societal health costs.

Recent articles:

T. Suna, A. Salminen, P. Soininen, R. Laatikainen, P. Ingman, M. Jauhiainen, M.-R. Taskinen, K. Kaski, M. Ala-Korpela: 1H NMR metabonomics of plasma lipoprotein subclasses – an elucidation of metabolic clustering by self-organising maps. NMR in Biomedicine, in press, 2006.

V.-P. Mäkinen, P. Soininen, C. Forsblom, M. Parkkonen, P. Ingman, K. Kaski, P.-H. Groop, M. Ala-Korpela: Diagnosing diabetic nephropathy by 1H NMR metabonomics of serum. Magnetic Resonance Materials in Physics, Biology and Medicine, in press, 2006.
A presentation by Ph.D. student Ville-Petteri Mäkinen based on this manuscript won the 2nd prise in the Young Investigator Award competition of the European Society for Magnetic Resonance in Medicine and Biology (ESMRMB) annual meeting in Warsaw, Poland, September 2006.

M. Ala-Korpela, N. Lankinen, A. Salminen, T. Suna, P. Soininen, R. Laatikainen, P. Ingman, M. Jauhiainen, M.-R. Taskinen, K. Héberger, K. Kaski: The inherent accuracy of 1H NMR spectroscopy to quantify plasma lipoproteins is subclass dependent. Atherosclerosis, in press, 2006.

T. Liimatainen, K. Lehtimäki, M. Ala-Korpela, J. Hakumäki: Identification of mobile cholesterol compounds in experimental gliomas by 1H MRS in vivo: effects of ganciclovir-induced apoptosis on lipids. FEBS Letters 580, 4746–4750, 2006.

M. Ala-Korpela, P. Sipola, K. Kaski: Molecular Detection and Characterisation of Atherothrombosis by magnetic resonance – potential tools for individual risk assessment and diagnostics. Annals of Medicine 38, 322-336, 2006.

K. Öörni, P. Posio, M. Ala-Korpela, M. Jauhiainen, P. T. Kovanen: Sphingomyelinase induces aggregation and fusion of small VLDL and IDL particles and increases their retention to human arterial proteoglycans. Arteriosclerosis, Thrombosis, and Vascular Biology 25, 1678-1683, 2005.

Selected background articles:

T. Väänänen, H. Koskela, Y. Hiltunen, M. Ala-Korpela: Application of quantitative artificial neural network analysis to 2D NMR spectra of hydrocarbon mixtures. Journal of Chemical Information and Computer Sciences 42, 1343-1346, 2002.

Š. Mierisová, M. Ala-Korpela: MR spectroscopy quantitation: a review of frequency domain methods. NMR in Biomedicine 14, 247-259, 2001.

M. T. Hyvönen, Y. Hiltunen, W. El-Deredy, T. Ojala, J. Vaara, P. T. Kovanen, M. Ala-Korpela: Application of self-organizing maps in conformational analysis of lipids. Journal of the American Chemical Society 123, 810-816, 2001.

K. K. Changani, R. Jalan, I. J. Cox, M. Ala-Korpela, K. Bhakoo, S. D. Taylor-Robinson, J. D. Bell: Evidence for altered hepatic gluconeogenesis in patients with cirrhosis using in vivo 31-phosphorus magnetic resonance spectroscopy. Gut 49, 557-564, 2001.

T. Hevonoja, M. O. Pentikäinen, M. T. Hyvönen, P. T. Kovanen, M. Ala-Korpela: Structure of low density lipoprotein (LDL) particles. Basis for understanding molecular changes in modified LDL. Biochimica Biophysica Acta - Molecular and Cell Biology of Lipids 1488, 189-210, 2000.

K. Öörni, M. O. Pentikäinen, M. Ala-Korpela, P. T. Kovanen: Aggregation, fusion, and vesicle formation of modified LDL particles: molecular mechanisms and effects on matrix interactions. Journal of Lipid Research 41, 1703-1714, 2000.

J. K. Hakala, K. Öörni, M. Ala-Korpela, P. T. Kovanen: Lipolytic modification of LDL by phospholipase A2 induces particle aggregation in the absence and fusion in the presence of heparin. Arteriosclerosis, Thrombosis, and Vascular Biology 19, 1276-1283, 1999.

K. Öörni, J. K. Hakala, A. Annila, M. Ala-Korpela, P. T. Kovanen: Sphingomyelinase induces aggregation and fusion, but phospholipase A2 only aggregation, of low density lipoprotein particles: two distinct mechanisms leading to increased binding strength of LDL to human aortic proteoglycans. Journal of Biological Chemistry 273, 29127-29134, 1998.

A. Korhonen, M. Jauhiainen, C. Ehnholm, P. T. Kovanen, M. Ala-Korpela: Remodeling of HDL by phospholipid transfer protein: demonstration of particle fusion by 1H NMR spectroscopy. Biochemical and Biophysical Research Communications 249, 910-916, 1998.

M. Ala-Korpela, M. O. Pentikäinen, A. Korhonen, T. Hevonoja, J. Lounila, P. T. Kovanen: Detection of low density lipoprotein particle fusion by proton nuclear magnetic resonance spectroscopy. Journal of Lipid Research 39, 1705-1712, 1998.

G. H. Haydon, R. Jalan, M. Ala-Korpela, Y. Hiltunen, J. Hanley, L. M. Jarvis, C. A. Ludlam, P. C. Hayes: Prediction of cirrhosis in patients with chronic hepatitis C infection by artificial neural network analysis of virus and clinical factors. Journal of Viral Hepatitis 5, 255-264, 1998.

J. Kaartinen, Y. Hiltunen, P. T. Kovanen, M. Ala-Korpela: Application of self-organising maps for the detection and classification of human blood plasma lipoprotein lipid profiles on the basis of 1H NMR spectroscopy data. NMR in Biomedicine 11, 168-176, 1998.

J. Rico-Sanz, J. V. Hajnal, E. L. Thomas, Š. Mierisová, M. Ala-Korpela, J. D. Bell: Intracellular and extracellular skeletal muscle triglyceride metabolism during alternating intensity exercise in humans. Journal of Physiology (London) 510, 615-622, 1998.

M. Ala-Korpela, P. Posio, S. Mattila, A. Korhonen, S. R. Williams: Absolute quantification of phospholipid metabolites in brain-tissue extracts by 1H NMR spectroscopy. Journal of Magnetic Resonance B 113, 184-189, 1996.

J.-P. Usenius, S. Tuohimetsä, P. Vainio, M. Ala-Korpela, Y. Hiltunen, R. A. Kauppinen: Automated classification of human brain tumours by neural network analysis using in vivo 1H magnetic resonance spectroscopic metabolite phenotypes. NeuroReport 7, 1597-1600, 1996.

M. Ala-Korpela, Y. Hiltunen, J. D. Bell: Quantification of biomedical NMR data using artificial neural network analysis: lipoprotein lipid profiles from 1H NMR data of human plasma. NMR in Biomedicine 8, 235-244, 1995.

M. Ala-Korpela: 1H NMR spectroscopy of human blood plasma. Progress in Nuclear Magnetic Resonance Spectroscopy 27, 475-554, 1995.

Y. Hiltunen, E. Heiniemi, M. Ala-Korpela: Lipoprotein-lipid quantification by neural-network analysis of 1H NMR data from human blood plasma. Journal of Magnetic Resonance B 106, 191-194, 1995.

M. Ala-Korpela, A. Korhonen, J. Keisala, S. Hörkkö, P. Korpi, L. P. Ingman, J. Jokisaari, M. J. Savolainen, Y. A. Kesäniemi: 1H NMR based quantitation of human lipoproteins and their lipid contents directly from plasma. Journal of Lipid Research 35, 2292-2304, 1994.

A. van den Boogaart, M. Ala-Korpela, J. Jokisaari, J. R. Griffiths: Time and frequency domain analysis of NMR data compared: an application to 1-D 1H spectra of lipoproteins. Magnetic Resonance in Medicine 31, 347-358, 1994.

M. Ala-Korpela, Y. Hiltunen, J. Jokisaari, S. Eskelinen, K. Kiviniitty, M. Savolainen, Y. A. Kesäniemi: A comparative study of 1H NMR lineshape fitting analyses and biochemical lipid analyses of the lipoprotein fractions VLDL, LDL, and HDL, and total human blood plasma. NMR in Biomedicine 6, 225-233, 1993.