学术报告厅

Metabonomics has been applied successfully in different research fields such as toxicology and pharmacology (1-7). However, application of metabonomics in nutritional research is still at infancy stage of development. Historically outcome measurement as an analytical approach has been widely used in nutritional research. This approach provides direct evidence linking the effect of nutrient intake on heath and disease. However, outcome measurement does not elucidate the underlying mechanistic effect of a nutrient intake in the biological system at cellular or molecular level. Thus, the link between a nutrient and its patho-physiological effect on the body is still missing. Metabonomics will shed light on understanding the mechanism of how nutrient intake affects health and disease.

In the previous paper, we have introduced the concept and potential application of metabonomics research in health sciences. The present paper will focus on the technical aspect of metabonomics in nutritional research. We will also overview the advantages and limitations of various laboratory analytical techniques currently available for metabolic profiling including experience from our laboratory in nutritional metabonomics research.

Pre-analytical precautions

Metabonomics is the analysis of small molecules that present in the biological samples except salts, protein and DNA. Successful application of metabonomics in nutritional research is not only dependent on how good instruments can perform but also the experimental design including procedures in sample handling and storage.

An expression of the phenotype of an individual is not only determined by the genetic makeup but the impact of the environment also contributes to the phenotypic variation. Metabolites are the “up-front” products of the “omics roadshow”. Thus, factors like gender, age, BMI, physiological and psychological status, physical activity, diurnal cycles, diet, gut microflora, etc. could potentially influence the metabolic profile (8-10). Therefore, a well controlled subject inclusion criteria is very important for a good study design. It has been demonstrated that consumption of a standard diet during the pre-study period reduced inter-individual metabolic profile variation in urine but not in plasma or salivary (9). Study in rodent has shown that different dietary intakes would affect serum metabolic profile in male and female rats (11). Because ingested nutrients will be absorbed and metabolised in the gastrointestinal tract and their metabolites will eventually enter into the body affecting metabolism and its metabolic profiles.

In sample handling and storage, special consideration must be taken to minimise the formation of artificial metabolites, adduction, and/or the degradation of metabolites. Acid is commonly used as a preservative to prevent bacteria growth in urine sample collection. However, the acid may hydrolyse the analytes in the samples and thus modify the metabolic profiles as such. Contamination of the sample container by bacteria will also modify the metabolites collected in the sample and ultimately affects the metabolite profiles. For serum collection, timing of serum collection, time allowed for clotting and the temperature for coagulation should be standardised in order to minimise sample variation.

Instrumentation and sample preparation

In metabonomics research, one of the biggest challenges is to measure, detect and quantify metabolites present in the biofluids. Different cutting edge technology have been developed for sample fractionation and/or quantitative mass detection. Nowadays, mass spectrometer (MS) and nuclear magnetic resonance (NMR) are the mainstream of analytical instruments for metabonomics research.

GC/MS

In the early 1970s, mass spectrometer was primarily used in molecular weight measurement. Nowadays, gas chromatography coupled with mass spectrometry (GC/MS) has become the most commonly used instrument for small molecules analysis (12-14). The combination of the high resolving capillary GC coupled with MS enables detection of small molecules such as amino acid and steroids for metabolic disorder diagnosis (15-17). Metabolic profiling of plant has also been demonstrated by using GC/MS (18).

GC/MS provides a high resolution analytical platform with a wide dynamic range (~106 orders of magnitude). The selection of the type and length of the GC column enhances the separation of particular molecules-of-interest which in turn increases the sensitivity for MS detection. However, sample preparation is the main disadvantage of using GC/MS in metabonomics. Sample extraction and derivatisation is usually required to improve the volatility of the sample. Although different methods of derivatisation are available, no single derivatisation step is capable of converting all metabolites present in biofluid. This is the biggest drawback of general metabolic profiling. Moreover, the relatively long analysis time also limits the number of samples to be analysed. Due to the aforementioned reasons, GC/MS has not been chosen as our analytical platform for general metabolic profiling.

NMR Spectroscopy

Nuclear Magnetic Resonance (NMR) spectroscopy was a popular analytical platform in 1990s. NMR relies on the quantum mechanical magnetic properties of the atom and nucleus. In brief, the electrons of the atom generate a small magnetic field which opposes to the external applied magnetic field. Chemically distinct molecule exhibits a unique chemical shift in the strong magnetic field and the signal is picked up by the detector inside the NMR spectrometer. Therefore, the stereochemical property of the molecule can be elucidated by analysing the unique spectra generated by NMR. In addition, the intensity of a particular signal also represents the concentration of the corresponding molecules presents in the biofluid. NMR was originally designed to determine chemical structure of molecules. In the past decade, NMR has been used for different metabonomics researches such as plant metabolism (19) and metabolic effects of flavonoid consumption in rats (20). Recently, the application of NMR has extended into clinical research such as the analysis of metabolic profiles of diabetes (21-22) and even in cancer researcher (23-24).

NMR spectroscopy is a promising technique which generates accurate information within a short period of time. The acquired information can also be used for molecule identification. Minimal sample preparation is also an advantage of using NMR. The recent development of multivariate data analysis tools such as Principle Component Analysis (PCA) further maximises the application of NMR in biomedical and clinical metabonomics research. Clustering of the chemical shift together with the change of intensities acquired from different samples provides a way to monitor the change of metabolic profile in complex biofluid. However, the major disadvantage in metabolic profiling is the poor dynamic range of NMR detection (~103 order of magnitude) and therefore only abundant metabolites are being detected and analysed.   Despite the poor dynamic range of the technique and the complexity of the data generated from a biological sample, NMR has been shown to be a high throughput analytical platform with a great potential in nutritional metabonomics (25).

LC/MS

Liquid chromatography coupled with mass spectrometer (LC/MS) has become a popular approach used in metabonomics. LC/MS provides an excellent analytical platform for metabolic profiling even for complex biological samples. The use of LC has been demonstrated to be an effective tool to separate the analytes from the complex mixtures according to their physical and/or chemical properties. Similar to GC, different columns provide a range of selectivity to interact with the analytes resulting in a high resolving power for particular analytes. Column and mobile phase selection in metabonomics is highly “result-dependent”. In general, to obtain a non-target metabolic profiling a reverse-phase column such as C-18 and a common organic solvent such as acetonitrile is commonly used.

Mass spectrometer coupled with and the ionisation mode has also improved the sensitivity, accuracy and dynamic range of metabolic profiling. Electrospray ionisation (ESI) is the most commonly used ionisation methods in MS. ESI, operates in both positive and negative ionisation mode, offers an excellent quantitative acquisition of different analytes present in a wide dynamic range with high reproducibility. It also provides a soft ionisation without fragmenting the analytes. The development of ESI facilitates the combination of LC and MS or even tandem MS (MS/MS), which provides an excellent analytical platform to separate the metabolites and thus to detect and quantify by MS. The separation of many metabolites by LC prior to the ESI has also been shown to minimise the ESI ion suppression and thus enhance the detection sensitivity of MS (26).

In our laboratory, Ultra Performance Liquid Chromatography coupled with Quadrupole Orthogonal Acceleration Time-of-Flight Mass Spectrometer (UPLC/QToF-MS) is used for metabonomics research. The UPLC system utilises the 1.7m porous particles and operates at a high pressure offering an improvement in speed, resolution and sensitivity. It has been demonstrated that UPLC is able to shorten the LC running time by one-third with much shaper peak when compare to the traditional HPLC method and more analytes can be detected (27-30). UPLC coupled to QToF-MS also offers a precise and accurate quantification of analytes. We have demonstrated the linearity of ions intensities of our instrument has a correlation of R2=0.97. The coefficient of variation (%CV) of inter-injection variation of the ion intensity is 13%. The retention time and the mass accuracy within 100 sample injections are less than 1% CV and 3 ppm respectively. All of these suggest the high reproducibility with good precision data acquisition makes UPLC/QToF-MS a high throughput metabolic profiling platform for different biofluids with confidence.

Human urine and serum are the most commonly analysed biofluids in our laboratory. They have been used for metabolic profiling in nutritional research such as bone metabolism and the health benefit of functional food. Urine is an aqueous-based biofluid which usually only contains trace amount of macromolecules and therefore filtration with 0.22m filter is usually sufficient to clean up the sample for LC/MS metabolic profiling. On the other hand, the analysis of serum metabolic profile involves a relatively complicated sample preparation steps prior to LC/MS metabolic profiling. Serum is rich in biomolecules including lipids, amino acids and proteins. Due to the size of the protein molecules and its possible reaction with the mobile phase, proteins precipitate out during LC and thus resulting in the blockage of the column especially in UPLC. The abundance of the protein will also cause ion suppression and affect the analysis of low abundant metabolites (31). Therefore, protein precipitation is required for serum sample prior to LC/MS analysis. Different methods and organic solvents have been evaluated for precipitating proteins in which organic solvents such as methanol has been shown to be the most effective protein precipitation agent with high reproducibility (31). Slow protein precipitation for 16 hours at 4oC has also been evaluated (32). The rationale behind the slow protein precipitation is due to the fact that serum protein, such as albumin, acts as a carrier in blood.　Fast protein precipitation would potentially remove the metabolites that are trapped inside albumin and lead to uncontrollable profiling results.　In addition, the storage condition of the extracted metabolites also affects the metabolite profiles.　We have shown that extracted metabolites are stable for up to 1 day regardless of the storage temperature (i.e. 4oC, -25C or -80oC) (unpublished data).　It again reinforces the importance of experimental planning especially in sample preparation step.　Nevertheless, sample preparation determines the metabolite profiles and therefore careful planning and selection of sample preparation procedures is crucial in metabolic profiling.

Challenge in data analysis and metabolite identification

The ultimate goal of nutritional metabonomics research is to identify and predict the metabolic pathways and effects of various nutrients on health and diseases. A huge dataset is generated by LC/MS and these data including the retention time, the mass-to-charge (m/z) and intensity which form a multi-dimensional matrix and specific bioinformatics tools are required to process these data. Different multivariate data analysis (MVDA) tools are available for clustering and classifying datasets (see previous paper). However, the chemical structure and the identity of the metabolite remain to be elucidated.

Accurate molecular mass and retention time may be sufficient to predict the identity of the metabolite; however, further analysis is required to confirm the identity. Although QToF-MS provides a highly accurate mass detection of less than 5 ppm, a study has shown that high mass accuracy even less than 1 ppm is not sufficient to determine the elemental composition of the metabolite particularly in high mass range (33). MS/MS provides further information of the parent ion based on the m/z of the fragmented ions but the identification of the metabolites largely depends on the MS/MS metabolite libraries. NMR provides an excellent platform to identify the molecular structure of the metabolite but purification step is required for accurate data interpretation.

Data processing is another challenge for metabolic profiling. The huge set of raw data points requires pre-processing before exporting into the multivariate data analysis tools. Data per-processing usually involves peak detection, peak matching, retention time alignment, baseline subtraction and normalisation (34-35). For example, one of the disadvantages of using HPLC/MS is the relatively poor resolution and the peak capacity that results in the shift of retention time. Internal standard may use to correct the retention time shift but the addition of chemical may mask or interference with the detection of endogenous metabolites. Furthermore, data pre-processing algorithms developed nowadays require, to a certain extend, the input of user-defined parameters and therefore an experienced data analyst is needed. An inappropriate data pre-processing parameters, such as retention time window, mass window and noise elimination level will affect the entire datasets and thus the clustering of metabolite profiles. Due to the aforementioned reasons, the classification and identification of metabolite as potential biomarkers remain the biggest challenge and further works is required.

Conclusion

Nutritional metabonomics in systems biology research explores the effect of nutrient intake on health and disease at cellular and molecular level. Potential biomarkers associated with the metabolic pathways pertaining to nutrient intake in health and disease could be identified. However, the technology has yet to be evolved to provide a high throughput analytical platform with high sensitivity and wide dynamic range to analyse the entire metabolome. The huge and complicated dataset, as well as the interactive roles of each metabolites in their own biochemical pathways make biomarker discovery in metabonomics even more difficult than other biomics to study. Therefore, the development of bioinformatic tools and metabonomics databases is important for elucidating the affected molecular pathways and the up- and down-regulation of specific metabolites in the complex biological system. Nevertheless, nutritional metabonomics is a powerful tool with great potential to develop into personalized nutrition.

Reference:

1.Filler, G. 2007, “Optimization of Immunosuppressive Drug Monitoring in Children”, Transplant.Proc., vol.39, no.4, pp.1241-1243.

2.Keun, H. C. 2006, “Metabonomic modeling of drug toxicity”, Pharmacol.Ther., vol. 109, no.1-2, pp.92-106.

3.Mazzucchelli, I., Franco, V., Fattore, C., Marchiselli, R., Perucca, E., & Gatti, G. 2007, “A Novel Enantioselective Microassay for the High-Performance Liquid Chromatography Determination of Oxcarbazepine and Its Active Metabolite Monohydroxycarbazepine in Human Plasma”, Ther.Drug Monit., vol.29, no.3, pp.319-324.

4.Nicholson, J. K., Connelly, J., Lindon, J. C., & Holmes, E. 2002, “Metabonomics: a platform for studying drug toxicity and gene function”, Nat.Rev.Drug Discov., vol.1, no.2, pp.153-161.

5.Plumb, R. S., Stumpf, C. L., Gorenstein, M. V., Castro-Perez, J. M., Dear, G. J., Anthony, M., Sweatman, B. C., Connor, S. C., & Haselden, J. N. 2002, “Metabonomics: the use of electrospray mass spectrometry coupled to reversed-phase liquid chromatography shows potential for the screening of rat urine in drug development”, Rapid Commun.Mass Spectrom., vol.16, no.20, pp.1991-1996.

6.Plumb, R. S., Stumpf, C. L., Granger, J. H., Castro-Perez, J., Haselden, J. N., & Dear, G. J. 2003, “Use of liquid chromatography/time-of-flight mass spectrometry and multivariate statistical analysis shows promise for the detection of drug metabolites in biological fluids”, Rapid Commun.Mass Spectrom., vol.17, no.23, pp.2632-2638.

7.Zhang, N., Fountain, S. T., Bi, H., & Rossi, D. T. 2000, “Quantification and rapid metabolite identification in drug discovery using API time-of-flight LC/MS”, Anal.Chem., vol.72, no.4, pp.800-806.

8.Lenz, E. M., Bright, J., Wilson, I. D., Hughes, A., Morrisson, J., Lindberg, H., & Lockton, A. 2004, “Metabonomics, dietary influences and cultural differences: a 1H NMR-based study of urine samples obtained from healthy British and Swedish subjects”, J.Pharm.Biomed.Anal., vol.36, no.4, pp.841-849.

9.Walsh, M. C., Brennan, L., Malthouse, J. P., Roche, H. M., & Gibney, M. J. 2006, “Effect of acute dietary standardization on the urinary, plasma, and salivary metabolomic profiles of healthy humans”, Am.J.Clin.Nutr., vol.84, no.3, pp.531-539.

10.Gibney M.J., Walsh M., Brennan L., Roche H.M., German B. and Ommen B.V. 2005, “Metabolomics in human nutrition: opportunities and challenges.” Am.J.Clin.Nutr., vol.82, no.3, pp.497-503.

11.Shi, H., Vigneau-Callahan, K. E., Shestopalov, A. I., Milbury, P. E., Matson, W. R., & Kristal, B. S. 2002, “Characterization of diet-dependent metabolic serotypes: primary validation of male and female serotypes in independent cohorts of rats”, J.Nutr., vol.132, no.5, pp.1039-1046.

12.Jansson, B. O. & Larsson, B. T. 1969, “Analysis of organic compounds in human breath by gas chromatography-mass spectrometry”, J.Lab Clin.Med., vol.74, no.6, pp.961-966.

13.Mamer, O. A., Crawhall, J. C., & Tjoa, S. S. 1971, “The identification of urinary acids by coupled gas chromatography-mass spectrometry”, Clin.Chim.Acta, vol.32, no.2, pp.171-184.

14.Viswanathan, C. V. 1974, “Coupled gas chromatography-mass spectrometry in the separation and characterization of polar lipids”, J.Chromatogr., vol.98, no.1, pp.105-128.

15.Deng, C., Deng, Y., Wang, B., & Yang, X. 2002, “Gas chromatography-mass spectrometry method for determination of phenylalanine and tyrosine in neonatal blood spots”, J.Chromatogr.B Analyt.Technol.Biomed.Life Sci., vol.780, no.2, pp.407-413.

16.Guillaume, C. P., van der Molen, J. C., Kerstens, M. N., Dullaart, R. P., & Wolthers, B. G. 1999, “Determination of urinary 18 beta-glycyrrhetinic acid by gas chromatography and its clinical application in man”, J.Chromatogr.B Biomed.Sci.Appl., vol.731, no.2, pp.323-334.

17.Wolthers, B. G. & Kraan, G. P. 1999, “Clinical applications of gas chromatography and gas chromatography-mass spectrometry of steroids”, J.Chromatogr.A, vol.843, no.1-2, pp. 247-274.

18.Kopka, J., Fernie, A., Weckwerth, W., Gibon, Y., & Stitt, M. 2004, “Metabolite profiling in plant biology: platforms and destinations”, Genome Biol., vol.5, no.6, pp.109.

19.Bligny, R. & Douce, R. 2001, “NMR and plant metabolism”, Curr.Opin.Plant Biol., vol. 4, no. 3, pp. 191-196.

20.Solanky, K. S., Bailey, N. J., Holmes, E., Lindon, J. C., Davis, A. L., Mulder, T. P., Van Duynhoven, J. P., & Nicholson, J. K. 2003, “NMR-based metabonomic studies on the biochemical effects of epicatechin in the rat”, J.Agric.Food Chem., vol.51, no.14, pp.4139-4145.

21.Griffin, J. L. & Nicholls, A. W. 2006, “Metabolomics as a functional genomic tool for understanding lipid dysfunction in diabetes, obesity and related disorders”, Pharmacogenomics., vol.7, no.7, pp.1095-1107.

22.Salek, R. M., Maguire, M. L., Bentley, E., Rubtsov, D. V., Hough, T., Cheeseman, M., Nunez, D., Sweatman, B. C., Haselden, J. N., Cox, R. D., Connor, S. C., & Griffin, J. L. 2007, “A metabolomic comparison of urinary changes in type 2 diabetes in mouse, rat, and human”, Physiol Genomics, vol.29, no.2, pp.99-108.

23.Griffin, J. L. & Kauppinen, R. A. 2007a, “A metabolomics perspective of human brain tumours”, FEBS J., vol.274, no.5, pp.1132-1139.

24.Griffin, J. L. & Kauppinen, R. A. 2007b, “Tumour metabolomics in animal models of human cancer”, J.Proteome.Res., vol.6, no.2, pp.498-505.

25.Puskas, L. G., Menesi, D., Feher, L. Z., & Kitajka, K. 2006, “High-throughput functional genomic methods to analyze the effects of dietary lipids”, Curr.Pharm.Biotechnol., vol.7, no.6, pp.525-529.

26.Matuszewski, B. K., Constanzer, M. L., & Chavez-Eng, C. M. 1998, “Matrix effect in quantitative LC/MS/MS analyses of biological fluids: a method for determination of finasteride in human plasma at picogram per milliliter concentrations”, Anal.Chem., vol.70, no.5, pp.882-889.

27.Nordstrom, A., O'Maille, G., Qin, C., & Siuzdak, G. 2006, “Nonlinear data alignment for UPLC-MS and HPLC-MS based metabolomics: quantitative analysis of endogenous and exogenous metabolites in human serum”, Anal.Chem., vol.78, no.10, pp.3289-3295.

28.Plumb, R., Castro-Perez, J., Granger, J., Beattie, I., Joncour, K., & Wright, A. 2004, “Ultra-performance liquid chromatography coupled to quadrupole-orthogonal time-of-flight mass spectrometry”, Rapid Commun.Mass Spectrom., vol.18, no.19, pp.2331-2337.

29.Wren, S. A. 2005, “Peak capacity in gradient ultra performance liquid chromatography (UPLC)”, J.Pharm.Biomed.Anal., vol.38, no.2, pp.337-343.

30.Zhao, X., Wang, W., Wang, J., Yang, J., & Xu, G. 2006, “Urinary profiling investigation of metabolites with cis-diol structure from cancer patients based on UPLC-MS and HPLC-MS as well as multivariate statistical analysis”, J.Sep.Sci., vol.29, no.16, pp.2444-2451.

31.Want, E. J., O'Maille, G., Smith, C. A., Brandon, T. R., Uritboonthai, W., Qin, C., Trauger, S. A., & Siuzdak, G. 2006, “Solvent-dependent metabolite distribution, clustering, and protein extraction for serum profiling with mass spectrometry”, Anal.Chem., vol.78, no.3, pp.743-752.

32.Boernsen, K. O., Gatzek, S., & Imbert, G. 2005, “Controlled protein precipitation in combination with chip-based nanospray infusion mass spectrometry. An approach for metabolomics profiling of plasma”, Anal.Chem., vol.77, no.22, pp.7255-7264.

33.Kind, T. & Fiehn, O. 2006, “Metabolomic database annotations via query of elemental compositions: mass accuracy is insufficient even at less than 1 ppm”, BMC.Bioinformatics., vol.7, pp.234.

34.Smith, C. A., Want, E. J., O'Maille, G., Abagyan, R., & Siuzdak, G. 2006, “XCMS: processing mass spectrometry data for metabolite profiling using nonlinear peak alignment, matching, and identification”, Anal.Chem., vol.78, no.3, pp.779-787.