Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
PDF hosted at the Radboud Repository of the Radboud University Nijmegen The following full text is a publisher's version. For additional information about this publication click this link. http://hdl.handle.net/2066/32101 Please be advised that this information was generated on 2017-06-18 and may be subject to change. NMR Spectroscopy of Body Fluids A Metabolomics Approach to Inborn Errors of Metabolism Udo Engelke NMR Spectroscopy of Body Fluids: A Metabolomics Approach to Inborn Errors of Metabolism Engelke, Udo Thesis Radboud University Nijmegen with a summary in Dutch ISBN 978-90-9022059-8 Cover Design by Robert Engelke Printed by Ponsen & Looijen BV, Wageningen NMR Spectroscopy of Body Fluids A Metabolomics Approach to Inborn Errors of Metabolism een wetenschappelijke proeve op het gebied van de Medische Wetenschappen Proefschrift ter verkrijging van de graad van doctor aan de Radboud Universiteit Nijmegen op gezag van de Rector Magnificus prof. mr. S.C.J.J. Kortmann, volgens besluit van het College van Decanen in het openbaar te verdedigen op donderdag 20 september 2007 om 10.30 uur precies door Udo Franziskus Heinrich Engelke geboren op 28 april 1966 te Kevelaer (Duitsland) Promotor Prof. dr. R.A. Wevers Copromotor Dr. M.A.A.P. Willemsen Manuscriptcommissie Prof. dr. A. Heerschap (voorzitter) Dr. M. Spraul (Bruker BioSpin GmbH, Duitsland) Dr. E. Morava Publication of this thesis was financially supported by - Department of Neurology, Radboud University Nijmegen Medical Center - Bruker BioSpin GmbH Voor mijn ouders CONTENTS 6 Abbreviations and symbols 8 Chapter 1 INTRODUCTION 11 1.1 General Introduction and aim of the thesis 13 1.2 NMR basics and Methods 17 1.3 NMR spectroscopy of body fluids as a metabolomics approach to inborn errors of metabolism. 27 N-ACETYLATED METABOLITES IN BODY FLUIDS 55 N-Acetylated metabolites in urine: proton nuclear magnetic 57 In: The Handbook of Metabonomics and Metabolomics. 2007; (14) 373 – 410. Published by Elsevier Editors: Lindon JC, Nicholson JK, Holmes E. Chapter 2 2.1 resonance spectroscopic study on patients with inborn errors of metabolism Clinical Chemistry 2004; 50(1): 58-66 Severe hypomyelination associated with increased levels of N-acetylaspartylglutamate in CSF 77 NMR spectroscopy of aminoacylase 1 deficiency, a novel inborn error of metabolism 91 Chapter 3 THE COMBINATION OF IN VIVO AND IN VITRO NMR SPECTROSCOPY 109 3.1 Dimethyl sulfone in human cerebrospinal fluid and blood plasma confirmed by one-dimensional 1H and twodimensional 1H-13C NMR 111 2.2 Neurology 2004; 62(9): 1503-1508 2.3 NMR in Biomedicine 2007; in press NMR in Biomedicine 2005; 18(5): 331-336 3.2 Methylsulfonylmethane (MSM) ingestion causes a significant resonance in proton magnetic resonance spectra of brain and cerebrospinal fluid 123 NMR spectroscopic studies on the late onset form of 3methylglutaconic aciduria, type I and other defects in leucine metabolism 129 Leukoencephalopathy in adult-onset 3-methylglutaconic aciduria type I 145 LIPID METABOLISM 155 Neuropediatrics 2006; 37(5): 312-314 3.3 NMR in Biomedicine 2006; 19(2): 271-278 3.4 Submitted for publication Chapter 4 4.1 Diagnosing inborn errors of lipid metabolism using NMR spectroscopy 1H- 157 Clinical Chemistry 2006; 52(7): 1395-1405 Chapter 5 SUMMARY AND FUTURE PERSPECTIVES 177 5.1 Summary and samenvatting 179 5.2 Future Perspectives 187 Reference List Dankwoord Curriculum Vitae List of Publications 195 211 213 215 Abbreviations and symbols δ γ υ0 1D 2D 3HIVA 3MG 3MGA 3MGH 7DHC 8DHC ACY1 ACY2 AUH B0 CNS COSY CSF CTX DMS DMSO DMSO2 EPI FID GC GC-MS GlcNAc HSQC IEM ISSD JRES LC-APCI-MS MAT I/III MRI MRS MRSI MSM NAA NAAG Chemical shift Gyromagnetic ratio Lamour precession frequency One-dimensional Two-dimensional 3-Hydroxyisovaleric acid 3-Methylglutaric acid 3-Methylglutaconic acid 3-Methylglutaconyl-CoA hydratase 7-Dehydrocholesterol 8-Dehydrocholesterol Aminoacylase type 1 Aminoacylase type 2 AU-specific RNA-binding enoyl-CoA hydratase Magnetic field strength Central nervous system Correlation spectroscopy Cerebrospinal fluid Cerebrotendinous xanthomatosis Dimethyl sulfide Dimethyl sulfoxide Dimethyl sulfone Echo planar imaging Free induction decay Gas chromatography Gas chromatography-mass spectroscopy N-acetylglucosamine Heteronuclear Single-Quantum Correlations Inborn error of metabolism Infantile sialic acid storage disorder J-resolved Liquid chromatography-atmospheric pressure chemical-ionization mass spectroscopy Methionine adenosyltransferase I/III Magnetic resonance imaging Magnetic resonances spectroscopy Magnetic resonance spectroscopic imaging Methylsulfonylmethane N-Acetylaspartic acid N-Acetylaspartylglutamic acid NMR OMIM OMS PLP PMD PPM PRESS SLOS T1 tCho tCr TGSE TLC TMS TOCSY TR TSP WET Nuclear magnetic resonance Online mendelian inheritance in man Octamethylcyclotetrasiloxane Proteolipid protein Pelizaeus-Merzbacher disease Parts per million Point resolved spectroscopy Smith-Lemli-Opitz syndrome Spin-lattice or longitudinal relaxation time Total Choline (choline, phosphocholine and glycerophosphocholine) Total Creatine (creatine and phosphocreatine) Turbo gradient spin echo Thin layer chromatography Tetramethylsilane Total correlation spectroscopy Repetition time Trimethylsilyl-2,2,3,3-tetradeuteropropionic acid Water suppression enhanced through T1 effects ONE INTRODUCTION 11 12 CHAPTER 1.1 General Introduction and aim of the thesis 13 14 Introduction and aim General introduction and aim of the thesis The precise number of different metabolites that play a role in human metabolism is unknown; estimates ranges from 3000 to around 20000 [1]. Approximately 30000 genes and 100000 proteins are involved in human cellular metabolism and homeostasis [1]. Inborn errors of metabolism (IEM) may occur in any enzymatic process that interconverts two metabolites. Currently more than 500 IEM are known [2]. These data suggest that many IEM remain as yet unknown. It seems that the analytical techniques that are used in routine diagnostics are insufficient to pick up a higher percentage of the estimated theoretical number of IEM. These techniques always rely on the detection of groups of metabolic intermediates with a specific chemical group. Because protons are available in almost any metabolite 1H-NMR spectroscopy may be able to simultaneously quantify almost all metabolites in the micromolar and millimolar concentration range. This opens the way towards a metabolomics view on metabolism. The aim of the present study is to explore the diagnostic potential of body fluid NMR spectroscopy. The outcome of the experiments can be the basis of a successful introduction of NMR spectroscopy as a complementary analytical technique in the metabolic laboratory of the (near) future. OUTLINE OF THE THESIS Chapter 1 of the thesis gives a brief introduction to the basic physical principles of NMR spectroscopy and discusses the diagnostic NMR approach, as developed in Nijmegen, for the diagnosis of inherited metabolic disorders (chapter 1.2). In chapter 1.3 we provide a literature overview of current applications of NMR spectroscopy to metabolic diseases. Chapter 2 describes the NMR findings for a group of compounds that would normally not be detected in a metabolic laboratory. These compounds contain an N-acetyl group and the study, as described in chapter 2.1, investigates thirteen known IEM involving N-acetylated metabolites. The diagnostic importance of Nacetylated metabolites is demonstrated in chapters 2.2 and 2.3: two novel diseases with accumulation of N-acetylated compounds in body fluid are described. Next to metabolomics of body fluids with 1H-NMR spectroscopy it is also possible to obtain metabolic information of tissues in humans by in vivo MR spectroscopy. In chapter 3 we used both in vivo and in vitro NMR spectroscopic techniques for metabolite identification. An unknown peak was identified as dimethyl sulfone (synonym: methylsulfonylmethane) in NMR spectra in human cerebrospinal fluid and blood plasma (chapter 3.1). The use of this compound as a dietary 15 Chapter 1.1 supplement led to the occurrence of an abnormal peak in the in vivo brain MR spectrum (chapter 3.2). In vivo and in vitro NMR studies on a patient with 3methylglutaconic aciduria type 1 are described in chapters 3.3 and 3.4. Chapter 4 investigates the potential use of 1H-NMR spectroscopy to identify and quantify lipids present in the blood of patients with different inborn errors of lipid metabolism. The technique provides a new tool in diagnosis and follow up of patients with inborn errors in lipid metabolism. 16 CHAPTER 1.2 NMR Basics and Methods 17 18 Basic principles of NMR spectroscopy Basic principles of NMR spectroscopy The clinical applications of nuclear magnetic resonance (NMR) are broad and diverse. In many hospitals, NMR is commonly used as a non-invasive technique to obtain clinical images (in vivo MRI, magnetic resonance imaging). In more specialized hospitals, the technique is used to obtain metabolic information of tissues in humans (in vivo MRS, magnetic resonance spectroscopy). Also, NMR has been used as an analytical technique to detect metabolites, drugs and toxic agents in body fluids (in vitro NMR). NMR EXPERIMENT The principle of NMR is based on the behavior of atoms placed in a strong magnetic field. Superconducting magnets provide such a strong, extremely homogenous magnetic field (Figure 1). Figure 1 Left: Standard NMR sample tube with 5 mm diameter. The sample volume needed is 650 µl. Middle: NMR spectrometer (500 MHz , 11.4 Tesla) with sample changer in the Nijmegen Goudsmit NMR pavilion. Right: NMR magnet operating at the world's highest, actively shielded field (950 MHz, 22.3 Tesla) The nuclei of most atoms (for example protons in a water sample) possess a property known as spin. The spinning of these charged particles generates a magnetic moment along the axis of spin, so that these nuclei act like tiny bar magnetics. Normally, the magnetic moments in a collection of nuclei will be randomly oriented. However, when an external magnetic field B0 is applied, they adopt one of a small number of allowed orientations with different energy. 1H nuclei can adopt only two permitted orientations, either aligned with (parallel) or against (anti-parallel) the direction of the external magnetic field. These two energy states are unequally populated, according to the Boltzmann distribution. The parallel orientation has a slightly lower energy state leading to more spin 19 Chapter 1.2 protons aligned parallel and leading to a net magnetization (Mz) (the vector sum of the individual magnetic moments). Figure 2 Figure 3 The collection of spins creating a net magnetization (Mz) A short radio-frequent radiation rotates Mz from (the vector sum of the individual magnetic moments). the z-axis into the x,y-plane. Mz will be aligned with the applied magnetic field (BO) which is defined as the z-direction in a x,y,z plane. This net magnetization will be aligned with the applied magnetic field (BO) which is defined as the z-direction in a x,y,z plane (Figure 2). In this equilibrium state, the net magnetization vector is static and does not induce a signal. To acquire an NMR signal, a short radio-frequent radiation (a pulse) is applied to the sample (Figure 3). This causes the magnetization vector Mz to rotate from the z-axis into the x,yplane (Figure 3). After a 90° pulse, the only remaining field is the external field (BO). Since (BO) and Mx,y are now perpendicular to each other, the magnetization vector experiences a torque, which causes it to precess around the magnetic field at the Larmor frequency (1): ν 0= ν0 B0 γ γ B0 2π (1) = Lamor precession frequency (Hz) = Magnetic field strength = Gyromagnetic ratio (different for each type of nuclei) The rotating magnetization induces a small oscillating electric current in a detection coil placed in the x,y-plane. The resulting signal, also called Free Induction Decay (FID), is a sum of many sine functions with different frequencies. 20 Basic principles of NMR spectroscpy Furthermore, the FID shows decay due to relaxation processes, which bring the magnetization vector back to its equilibrium state. The FID can be converted to a frequency-domain NMR spectrum by Fourier transformation. PARAMETERS FROM NMR MEASUREMENTS A 1H-NMR spectrum provides a characteristic ‘fingerprint’ of almost all protoncontaining metabolites. The spectral parameters chemical shift (δ in ppm), spinspin coupling (multiplicity) and signal intensity are important for body fluid analysis and will be described briefly here. Chemical Shift (Resonance position) The Larmor frequency is determined by the gyromagnetic ratio and the strength of the external field BO (equation 1, page 20). However, the resonance frequency also depends on the chemical environment, since nearby nuclei and electrons slightly alter the local magnetic field of a specific proton. This effect, called the chemical shift, causes nuclei with different chemical environments to resonate at different frequencies and therefore provides structural information. The position of a frequency, in absolute frequency terms, varies with the exact field strength of the magnet used. Therefore, a chemical shift reference is used to define the positions of the resonances in the spectrum. These shifts are generally very small and are defined as parts per million (ppm) fractional shift from an internal reference compound according to equation 2. ⎛ ⎞ 6 δ = ⎜⎜ Shift from an internal reference in Hz ⎟⎟ ×10 ⎝ Spectromet er frequency in Hz (2) ⎠ δ = Chemical shift (ppm) The chemical shift expressed as ppm is not dependent on the field strength of the external magnetic field. For body fluid NMR spectroscopy, trimethylsilyl-2,2,3,3tetradeuteropropionic acid (TSP) is often used as a chemical-shift reference. The chemical shift can be used to discriminate the 1H-NMR spectra of molecules, even when their chemical structure is only slightly different. For example, two molecules that have a quite similar chemical structure are lactic acid and alanine. The only difference is that lactic acid has a hydroxyl group whereas alanine has an amino group. As shown in Figure 3, the CH3-protons of lactic acid resonate at 1.41 ppm and the CH-proton resonates at 4.37 ppm under the conditions used. 21 Chapter 1.2 Due to the small difference in chemical structure, the resonance positions of the CH3- and CH-protons of alanine are slightly different (the CH3-protons resonate at 1.50 ppm, the CH proton at 3.89 ppm). Spin-spin coupling (Multiplicity) NMR spectra furthermore provide valuable information about the magnetic interactions between nuclei (also known as spin-spin coupling, scalar coupling or J-coupling). These couplings cause the signal of each nucleus to split into a distinct number of compounds with characteristic relative intensities and spacings, which depend on the number and type of spins coupling with the nucleus of interest. This is illustrated using the spectrum of lactic acid shown in Figure 3. The splitting of the resonances is caused by an interaction between neighbouring protons. Some rules governing this splitting are: 1. No splitting is caused between equivalent protons, e.g. the CH3-group protons of lactic acid. 2. A proton that is coupled to n equivalent protons gives rise to (n + 1) lines. The relative intensities of these lines are given by the binomial distribution. In Figure 3, the equivalent CH3-group protons coupled to the CH-group proton give rise to two lines (a doublet) with relative intensities of 1:1. The CH-group proton coupled to the equivalent CH3-group protons gives rise to four lines (a quartet) with relative intensities of 1:3:3:1. Since the hydroxyl proton in lactic acid exchanges rapidly with water protons under the conditions used, it does not couple to any of the nonexchangeable protons in this molecule. 22 Basic principles of NMR spectroscpy Figure 3 1H-NMR spectra of lactic acid (upper panel) and alanine (lower panel) dissolved in D2O measured at 500 MHz; pD = 2.5 Signal intensity The peak area or signal intensity of a resonance in a NMR spectrum is proportional to the number of protons contributing to the signal when appropriate experimental conditions are used. For example in Figure 3, the doublet is assigned to the CH3-group (3 protons contributing) and the quartet is assigned to the CHgroup (1 proton contributing). Therefore, the peak area of the doublet is three times as large as the peak area of the quartet. Since the peak area is proportional to the number of protons contributing to the signal, it is also proportional to the concentration of the molecule of interest. Therefore, it is possible to use NMR spectroscopy for metabolite quantification. The detection limit of the technique is in the low micromolar range for most metabolites. NMR SPECTROSCOPY OF BODY FLUIDS IN PATIENTS WITH INBORN ERRORS OF METABOLISM In 1992, Prof. Dr. Ron Wevers, started the work on NMR spectroscopy for diagnosis and follow-up of patients with inborn errors of metabolism in Nijmegen. During the subsequent years, this approach was successful and standard methods for recording 1H-NMR spectra in cerebrospinal fluid and blood plasma have been published [3;4]. 23 Chapter 1.2 Sample preparation is limited to the removal of proteins from CSF and plasma serum samples on a 10 kD filter and to pH standardization of all samples (pH 2.5). Routinely we start with a 1D 1H-NMR experiment. By comparing the spectrum with databases containing model compounds and normal body fluid NMR spectra, assignments can be made. This may lead to the diagnosis of an IEM [5]. Sometimes the overlap of resonances is too severe and assignments are not possible from a 1D spectrum. In such case a two-dimensional COSY NMR experiment is used to provide additional information (Figure 3B, page 36). Sometimes unknown resonances are not present in the database. If we can exclude medication or food component as source of these unknown peaks additional NMR experiments can be used to identify the structure of the unknown metabolite (Table 1, page 34). Which body fluid should be investigated? It is difficult to give a general advice on the type of body fluid that is most suited to find the diagnosis. Urine and blood are easy to collect and urine is often used as a first approach. The kidneys rapidly excrete some metabolites but others remain preferentially in the blood. Urine NMR spectra are very complex and many IEM can be diagnosed in this body fluid [5]. CSF NMR investigation may be valuable for patients suspected of suffering from neurometabolic disease. Chapter 2.2 demonstrates the presence of the dipeptide N-acetylaspartylglutamic acid (NAAG) in CSF of two patients with a novel neurometabolic disease while NAAG could not be demonstrated in the urine NMR spectra of the patients. Case selection NMR spectroscopy of body fluids may be considered when the patient is suspected of suffering from a so far undiagnosed inborn error of metabolism or when further investigations leading to a diagnosis are required. Clinical indications are: 1. Two or more children in the same family have unexplained similar clinical signs and symptoms. 2. The patient has an unusual body odor 3. An unknown metabolite is observed using in vivo NMR spectroscopy of the brain or other organ (Chapter 3). Biochemical indications are: 1. An abnormal unknown metabolite is repeatedly observed in body fluids with another technique. NMR spectroscopy may provide structural information about the metabolites involved. 2. Confirmation of a special diagnosis with an independent technique is needed (Chapter 2.3). 24 Basic principles of NMR spectroscpy 3. The need for a reliable quantification of a metabolite that is difficult to quantify otherwise (Chapter 2 and 3). 4. Consistent abnormal results were found using conventional screening techniques for one or more metabolites that did not lead to a classifying diagnosis. NMR spectroscopy may provide additional information. 25 26 CHAPTER 1.3 NMR spectroscopy of body fluids as a metabolomics approach to inborn errors of metabolism Udo F.H. Engelke, Marlies Oostendorp and Ron A. Wevers In: The Handbook of Metabonomics and Metabolomics. 2007; (14) 373 – 410. Published by Elsevier Editors: Lindon JC, Nicholson JK, Holmes E 27 28 NMR spectroscopy of body fluids Introduction Nuclear magnetic resonance (NMR) spectroscopy has been used to analyze metabolites, drugs and toxic agents in body fluids and reviews on these topics have been published [6-8]. Proton NMR (1H-NMR) spectroscopy has also been successfully applied to the field of inborn errors of metabolism [9]. The technique is essentially non-selective and provides an overview of protoncontaining metabolites. Moreover, simultaneous quantification of many metabolites over large concentration ranges can be done easily in a short time-frame. Because of these features proton NMR spectroscopy easily picks up a metabolic derangement or an abnormality in the metabolome. The technique either as a standalone approach or in combination with liquid chromatography and/or mass spectroscopy is at the frontiers of metabolomics and systems biology approaches. THE DIAGNOSTIC APPROACH OF INBORN ERRORS OF METABOLISM Inborn errors of metabolism (IEM) form a considerable group of genetic diseases. The majority is due to defects of single genes coding for enzymes. Figure 1A shows a schematic representation of a metabolic pathway. In most of the disorders, problems arise due to accumulation of substances, which are toxic or interfere with normal function (Figure 1B, metabolites S1 and S2). Clinical signs may also relate to the decreased availability of metabolites behind the metabolic block (Figure 1B, metabolites S3 and S4). Also alternative pathways may metabolize an accumulating metabolite. Such pathways may play a minor role or no role at all in normal metabolism. In a patient with a metabolic block alternative pathways may result in unusual metabolites (Figure 1B, metabolites S5 and S6). These often are characteristic for the disease and therefore diagnostically important. Also they may play a role in the pathophysiology of the disease. Figure 1 A; A schematic representation of a metabolic pathway. Enzymes E1, E2 and E3 catalyze the reactions converting the metabolite S1 into S2, S3 and S4. B; A defect in enzyme E2 (black square) will give rise to the accumulation of metabolites S2 and to a lesser extent S1 and to a decreased availability of metabolites S3 and S4. Also unusual alternative metabolites like S5 and S6 may occur. 29 Chapter 1.3 One of the methods to diagnose IEM is the detection of abnormal metabolite concentrations in body fluids such as urine, blood plasma, blood serum or cerebrospinal fluid (CSF). From these concepts it is easy to understand that the diagnosis of an IEM may rely on (1) an abnormal accumulation of a specific metabolite, (2) the presence of metabolites that are normally not present in body fluids or (3) a decreased concentration of a metabolite that is always present in the body fluid. Laboratories specialized in diagnosing IEM use a variety of analytical methods to measure various metabolic groups. These are for instance GC-MS for the detection of organic acids, ion exchange chromatography for the detection of amino acids and HPLC for the detection of purines and pyrimidines. These techniques are selective and sample pretreatment such as derivatization and extraction procedures are often necessary. Proton nuclear magnetic resonance (1H-NMR) spectroscopy is not selective and because most metabolites contain protons this technique may be considered as an alternative analytical approach for diagnosing known, but also as yet unknown, IEM. The detection limit for body fluid NMR spectroscopy depends on a number of factors such as field strength, number of protons contributing to a resonance and the region of the spectrum where the resonance is observed. In general, the detection limit is in the low micromolar range in the less crowded regions of the spectrum. For example, the detection limit at S/N = 3 for the six equivalent protons in dimethyl sulfone (observed as a singlet resonance at 3.14 ppm in blood plasma and CSF) was determined to be 0.5 µmol/l (500 MHz, 128 transients; using a conventional 5 mm TXI probehead operating at ambient temperature) [10]. Another study showed detection limts between 2 and 40 µmol/l for various metabolites in unconcentrated plasma samples measured at 600 MHz and 132 transients [4]. Metabolite quantification in plasma and CSF can be performed by adding a known concentration of 3-(trimethylsilyl)-2,2,3,3-tetradeuteropropionic acid (TSP) to the sample and concentrations can be expressed as µmol/liter. For urine, the singlet resonance of creatinine at 3.13 ppm (pH=2.50) can be used as a concentration reference. Metabolite concentrations in urine are often expressed per mmol creatinine. For several compounds good correlations have been obtained with other analytical techniques. In CSF and blood plasma, quantitative data for alanine, threonine and lactic acid correlated well with data obtained with conventional techniques [3;4]. Excellent linearity, recovery and reproducibility have been found for N-acetyltyrosine, N-acetyltryptophan and N-acetylaspartic acid in urine, normally not analyzed in conventional screening methods [11]. In this review, we provide a literature overview of current applications of NMR spectroscopy to metabolic disease. 30 NMR spectroscopy of body fluids Methods SAMPLE COLLECTION, PREPARATION AND CHOICE NMR spectroscopy has been used to investigate a wide range of body fluids [7]. However, when used as a diagnostic tool for inborn errors of metabolism, NMR spectra are mainly recorded on blood plasma, blood serum, urine or CSF samples. Blood is collected by venipuncture into standard vials containing either lithium heparin or ethylene diamine tetra acetate (EDTA) as anti-coagulant. Subsequently, plasma can be separated from the cells by centrifugation. When EDTA is used to prevent clotting, extra resonances can be observed in the NMR spectrum due to complexes formed between EDTA and Ca2+ and Mg2+ ions present in plasma [12]. CSF can be obtained by lumbar puncture and has to be centrifuged to remove cells. If NMR analysis cannot be performed immediately after collection of blood, urine or CSF, the samples have to be kept frozen at -20°C or lower until further usage. Whereas urine is measured directly, plasma and serum are usually deproteinized. The blood plasma protein concentration (approximately 70 g/l) results in broad overlapping signals that obscure resonances from low- molecular mass metabolites of interest [4]. Adsorption of the commonly used internal standard trimethylsilyl-2,2,3,3-tetradeuteropropionic acid (TSP-d4) on proteins has been shown to have a negative influence on quantification [13]. Additionally, resonances for small, protein-bound metabolites are broadened due to T2-relaxation processes, making accurate quantification difficult [14]. The effect of deproteinization on the 1H NMR spectrum of a control plasma sample is illustrated in Figure 2. Normal CSF has a rather low protein concentration (< 500 mg/l) and NMR measurements on both native [15-20] and deproteinized samples have been reported [3]. Various techniques are available for plasma or serum deproteinization, including extraction with acetone or acetonitrile, perchloric acid precipitation, solid phase extraction chromatography and ultrafiltration. Of these methods, ultrafiltration using a 10 kDa cut-off filter (e.g. Sartorius) is best suited for metabolic studies [21]. Before use, the filter has to be washed by centrifugation of water to remove glycerol that is present in several commercially available filters. The main disadvantage of deproteinization is possible loss of protein bound metabolites. However, excellent correlation between 1H-NMR and enzymatically determined lactate concentrations was found [4], even though lactate is assumed to interact with plasma proteins and so become partly NMR invisible [22]. The pH of the sample has an important influence on the observed chemical shifts and also determines whether acidic or basic protons are detectable. Unfortunately, a large range of pH values has been used in the literature. The physiological pH and pH of 2.5 have been used most frequently. The pH of urine varies considerably between samples. The pH of CSF samples increases significantly upon standing. To 31 Chapter 1.3 A B Figure 2 500 MHz 1H-NMR spectrum of a plasma sample before (B) and after (A) deproteinization over a 10-kilodalton filter. Assignmets: Isoleucine (Ile), leucine (Leu), valine (Val), alanine (Ala), lysine (Lys), glutamine (Glu), tyrosine (Tyr), phenylalanine (Phe), histidine (His). HDL Chol referes to the C18 signal from cholesterol in high-density lipoprotein. CH3 and CH2 refer to the CH3 and CH2 groups of fatty acids in lipoproteins. L1-L6 refer to proton signals of various lipids [23]. 32 NMR spectroscopy of body fluids improve intersample reproducibility and to allow comparison between different samples, standardization of pH was initially proposed by Lehnert and Hunkler in 1986 [24]. They chose pH 2.5 because the slope of the titration curve is expected to be minimal for all organic acids and because the proton concentration is not yet so high as to cause hydrolysis of certain compounds. Chemical shifts in this review are given for pH 2.5 unless mentioned otherwise. Finally, an aliquot of TSP-d4 in 2H2O has to be added to the sample to provide a chemical shift (δ = 0.00) and concentration reference as well as a deuterium lock signal. Alternatively, an internal reference or the residual solvent signal can be used as chemical shift reference. In this case addition of pure 2H2O is sufficient. The ionic strength of body fluids varies considerably and may be high enough to adversely affect the tuning and matching of the RF circuits of a spectrometer probe, particularly at high field strengths. Therefore, it may be necessary to retune the spectrometer probe for each sample. Automatic tune and match (ATM) probes are available. Traditionally the approach in diagnosing IEM is to make use of urine as the body fluid of first choice. The kidneys excrete most water-soluble small molecules. However, some are reabsorbed like for instance glucose. The urine NMR spectrum is by far the most complex body fluid spectrum allowing the diagnosis of most inborn errors of metabolism. However, there may be a case for separate CSF NMR studies in patients suspected to suffer from neurometabolic disease. A paper by Wolf et al. [25] demonstrated the presence of the dipeptide Nacetylaspartylglutamic acid (NAAG) in CSF of two patients with a novel neurometabolic disease while NAAG could not be demonstrated in the urine NMR spectra of the patients. Only single studies have addressed the NMR spectra of body fluid cells. Studies are available in leucocytes and erythrocytes [26]. The composition of the intracellular fluid differs considerably from the body fluid composition. Several metabolites are unable to leave the cell, as they cannot pass the outer cell membrane. Phosphorylated compounds may serve as an example. Inborn errors where such compounds accumulate may be easily diagnosed in homogenates of leucocytes. However, such examples have not yet been presented in literature. ONE AND TWO-DIMENSIONAL NMR SPECTROSCOPY TECHNIQUES The number of observed metabolites in body fluids is largely dependent on the magnetic field strength of the NMR spectrometer. At higher fields, the spectral dispersion and sensitivity increase, allowing assignment of many metabolites that cannot be detected at lower fields. Therefore, working at the highest field available will provide the most complete metabolic information. Currently, most studies are performed on 400, 500 or 600 MHz spectrometers, but higher field strengths up to 800 MHz have been used [23;27;28]. Additionally, low frequency spinning of 33 Chapter 1.3 the sample can be used to improve spectral resolution. Reported measurement temperatures are found between 20 and 37˚C. Table 1: Overview of different experiments used in the diagnosis of inborn errors of metabolism. Experiment Description 1D 1H (Figure 3A) Single pulse with solvent presaturation. Routinely used to determine chemical shifts, J-values and metabolite quantities (only for well-resolved peaks). H-1H COSY (Figure 3B) Used to establish which protons are spin-coupled. Helpful in resolving H-1H DQF-COSY COSY variant with greater resolution. Cross peaks near the diagonal can be H-1H TOCSY (Figure 3C) COSY variant displaying relayed connectivities. Used as an aid in assignment. H J-resolved (Figure 3F) Separates J-splitting and chemical shifts on to two orthogonal axes. 1 1 overlap and to assign crowded regions. resolved. Sensitivity is 50% less compared to COSY. 1 1 Reduced overlap and more accurate determination of J-values and chemical shifts. Spectral simplification based on difference in relaxation times. Not suitable H 1D Spin-echo 1 1D 1H-decoupled for quantification. C (Figure 3D) More resonances resolved due to larger spectral width than 1H. Suffers from inherent low sensitivity and low natural abundance. 13 H-13C HSQC (Figure 3E) 1 Displays connectivities between chemically bounded Helpful when assigning complicated spectra. H and 1 C nuclei. 13 A wide range of one-dimensional (1D), two-dimensional (2D) and even higher dimensional NMR experiments are available to the modern NMR operator. Especially in the field of structural biology, multidimensional experiments are important for successful determination of protein or nucleic acid structures [29;30]. For studying inborn errors of metabolism, 1D 1H NMR spectra are most frequently employed. Additionally, several homo- and heteronuclear 2D experiments can be recorded, which are particularly helpful in assisting the assignment of resonances. The most commonly used experiments for diagnosing metabolic diseases will now be discussed. The information content of the spectra will be demonstrated using spectra recorded on a sample containing pure methionine (Figure 3). A schematic overview of the experiments is in Table 1. ONE-DIMENSIONAL (1D) NMR SPECTROSCOPY Since proton NMR is relatively sensitive and because protons are present in almost every metabolite, a 1D, single pulse 1H NMR experiment with solvent presaturation 34 NMR spectroscopy of body fluids is a good starting point when diagnosing a possible inborn error of metabolism. When overlap is not too severe and metabolite concentrations are high enough (> 2 – 40 µmol/l), the chemical shifts and J-coupling constants can be determined directly and subsequent quantification of these metabolites is possible. The singlet resonance at 2.13 ppm from the ε methyl protons of methionine is most suited to quantify this metabolite in body fluids (Figure 3A). However, assignment can be rather difficult when high order spin coupling patterns are present or when spectral regions are crowded. Furthermore, presaturation of the water signal can result in loss of nearby peaks. TWO-DIMENSIONAL (2D) NMR SPECTROSCOPY Two-dimensional NMR experiments can provide additional information to solve overlap problems and allow identification of metabolites that otherwise remain undetected. They are based on the couplings between magnetic nuclei. These can be dipolar (through space) or scalar (through bond) couplings. Here, only experiments from the latter category will be discussed. 1H-1H correlation spectroscopy (COSY) The first homonuclear 2D experiment is correlation spectroscopy (COSY). The underlying concept is coherence transfer from one spin to another via J-coupling [31]. It reveals the network of spin-spin couplings in each molecule and can therefore be used to aid in spectral assignment. The peaks under the diagonal of a COSY spectrum correspond to the peaks normally observed in a 1D spectrum. The off-diagonal peaks (cross peaks) provide the interesting information on which protons are spin coupled. From the COSY spectrum of methionine for instance it is clear that the α -proton is J-coupled to the β-protons (Figure 3B). Unfortunately, there is a 90˚ phase difference between the diagonal and cross peaks, which makes it impossible to correct the phase of a COSY spectrum in such a way that all peaks have absorption phase. When the phase of the cross peaks is corrected to pure absorption, the diagonal peaks display dispersive line shapes in both dimensions and their long dispersion tails can conceal nearby cross peaks. A solution is to record a Double Quantum Filtered COSY (DQF-COSY) experiment, in which both diagonal and cross peaks have absorption line shapes. Consequently, the spectral resolution is improved and cross peaks close to the diagonal can be readily detected. 35 Chapter 1.3 Figure 3 Six different 500 MHz NMR spectra of methionine. (Legends page 37) 36 NMR spectroscopy of body fluids Figure 3 (Page 36) Six different 500 MHz NMR spectra of methionine. A: 1D 1H-NMR spectrum ε ε S-C H3: The methyl protons on C give a singlet resonance at 2.13 ppm. This singlet corresponds to the three equivalent protons of the methyl group. The fact that it is a singlet illustrates that the neighboring atom (the S atom) does not carry any proton(s). In body fluids the 2.13 ppm resonance is often used to quantify methionine. γ β γ C H2: Due to the J-coupling with the two neighboring protons on C the protons on C give a triplet resonance γ at 2.63 ppm ( Jγ, β = 7.3 Hz) β β α C H2: The protons on C give two multiplets at 2.10 and 2.20 ppm. This is explained by the fact that C is an β asymmetrical carbon atom, and the two methylene protons on C are therefore diastereotopic; consequently α γ they give separate resonances. Due to the J-coupling with the proton on C and the protons on C we find two multiplets. α α C H: Due to the J-coupling with the two diastereotopic methylene protons on Cβ, the proton on C gives a disturbed triplet (doublet/doublet) resonance at 3.95ppm. The peak area for the various resonances in methionine is proportional to the number of protons attached to the various C atoms. B: 2D 1H-1H COSY spectrum Chemical shift and scalar coupling can be obtained simultaneously for methionine by detecting the offdiagonal cross peaks, which are symmetric with respect to the diagonal. The diagonal and cross peaks connected by dashed lines indicate which protons of methionine have a scalar coupling. The diagonal peak of β the two protons on C forms a corner of two squares, as these protons are coupled both to the proton on C α γ and to those on C . C: 2D 1H-1H TOCSY spectrum Compared with the COSY spectrum (Figure 3B) two new signals have now appeared. These provide evidence of α γ a correlation between the protons on C and C . D: 1D 1H-decoupled 13 C spectrum In the 1D 1H-decoupled α β γ ε C NMR spectrum, C , C , C and C from methionine give resonances at 56.7, 32.5, 13 31.7 and 16.7 ppm, respectively. Note: This 1D 1H-decoupled 13 C spectrum was recorded at pH 7.0. E: 2D 1H-13C HSQC spectrum This type of NMR spectroscopy reveals which carbons and protons in methionine are chemically bonded. For methionine, the 1H/13C shift values for the correlation peaks are 3.86 / 56.7 (α), 2.20; 2.10 / 32.5 (β), 2.65 / 34.7 (γ) and 2.13 / 16.7 (α). Note: This HSQC spectrum of methionine was recorded at pH 7.0. F: 2D 1H J-resolved spectrum The J-multiplets are displayed along the F1 dimension, while the chemical shifts are shown in the F2 dimension. This minimizes overlap and allows better determination of coupling constants, which also are detectable in 1D γ γ 1H-NMR. The protons on C give a triplet resonance at 2.63 ppm with a J-coupling of 7.3 Hz ( J γ, β = 7.3 Hz) Also, chemical shifts and coupling constants can be determined more easily. However, a disadvantage is that the sensitivity is reduced compared to a regular COSY, because only two out of four scans lead to cross peaks and the other two lead to diagonal peaks [31]. Another solution to the phase problem is displaying 37 Chapter 1.3 the conventional COSY spectrum in magnitude mode. All peaks appear absorptive, although cross peaks near the diagonal are still likely to be obscured. In literature, COSY spectra of body fluids have been widely used in patients with IEM, e.g., in fucosidosis [11], 3-hydroxy-3-methylglutaryl CoA lyase deficiency [32], ribose 5phosphate isomerase deficiency [33] and Salla disease [11;34]. Total Correlation Spectroscopy (TOCSY) COSY spectra of complex mixtures like body fluids can still show a substantial amount of overlap. Resulting ambiguities in the spectral assignment may be resolved by detecting multiple relayed connectivities using a Total Correlation Spectroscopy (TOCSY) experiment. For instance, in an AMX spin system cross peaks are observed between A and X, where both spins are coupled to spin M, but not to each other. This allows detection of long range interactions (>3 bonds), that are usually too weak in normal COSY spectra. Other advantages are a higher sensitivity for larger molecules and absorption line shapes for both diagonal and cross peaks. The number of observed cross peaks depends on the length of the applied TOCSY mixing sequence. A spectrum recorded with a short mixing period mainly contains cross peaks of directly coupled spins, whereas long mixing gives rise to relay cross peaks. In the case of methionine, the TOCSY spectrum (Figure 3C) for instance shows cross peaks between the α-proton and the γ-protons, which were not present in the regular COSY (c.f. Figure 3B). 750 MHz 1D TOCSY spectra from seminal fluid was published by Spraul et al. [35]. Furthermore, 1D TOCSY has been used to identify 3-ureidoisobutyric acid as accumulating compound in ureidopropionase deficiency [36]. To our knowledge there are no examples of successful application of 2D TOCSY to the field of inborn errors of metabolism. J-resolved spectroscopy (JRES) Another useful homonuclear 2D experiment is the J-resolved (JRES) experiment, which separates the J fine structure from the chemical shifts. In the resulting spectrum, the J-multiplets are displayed along the t1 dimension, while the chemical shifts are shown in the t2 dimension. This minimizes overlap and allows better determination of coupling constants. Furthermore, a proton decoupled 1D spectrum can be obtained by projection of the 2D J-resolved spectrum onto the F2 axis. The JRES spectrum of methionine is shown in Figure 3F. 600 MHz 2D JRES 1H-NMR spectra from urine and blood plasma samples was published by Foxall et al. [37]. In literature, JRES 1H-NMR spectra of body fluids have been used in patients with IEM, e.g, ribose 5-phosphate isomerase deficiency [33], maple syrup urine disease [27] and 2-methylacetoacetyl CoA thiolase deficiency [38]. 38 NMR spectroscopy of body fluids Spin-echo spectra Proton spectra can be also simplified by making use of the differences in relaxation behavior between different molecules. This principle forms the basis of spin-echo experiments. In spin-echo spectra, signals arising from relatively large molecules (i.e.protein signals in intact plasma) are attenuated due to their shorter relaxation times. Furthermore, even-numbered multiplets are displayed as negative peaks, whereas singlets and odd-numbered multiplets appear positive [38;39]. Spin-echo experiments can therefore be helpful when assigning complex spectra. However, the method is less suited for quantification due to signal loss during long echo times. Since plasma samples are routinely deproteinized, spin echo experiments are not recorded often nowadays when trying to diagnose an inborn error of metabolism. 13C NMR Spectroscopy With the advances made in NMR spectrometer technology, 13C NMR becomes more and more feasible. Despite its low intrinsic sensitivity (~60 times less than 1H) and low natural abundance (1.1%), 13C NMR can provide a wealth of additional information. This is mainly because of the large range of observed chemical shifts, resulting in many resolved signals, and because precursors of the metabolite of interest can be isotopically enriched. This last option enhances sensitivity and can simplify the study of metabolic pathways [40]. Another advantage of 13C NMR spectroscopy is the absence of the water resonance, making solvent suppression no longer necessary. The simplest 13C method used in the field of inborn errors of metabolism is the 1D 1H decoupled 13C experiment. Proton decoupling results in enhanced sensitivity due to collapsing of multiplets into singlets and due to the Nuclear Overhauser Effect (NOE; up to 4 times for 13C). By comparing the decoupled spectrum with the corresponding 1H-coupled spectrum, structural information can be obtained about the number of protons bound to each carbon [40]. As an example, the 1D proton decoupled 13C spectrum of methionine is shown in Figure 3D. Although 1D 13C spectra are not commonly recorded yet, their effective use in studying several inborn errors of metabolism, including galactosemia [41], dimethylglycine dehydrogenase deficiency [42], 2-hydroxyglutaric aciduria [43], Canavan disease [44;45] and argininosuccinic aciduria [46], has been shown. Heteronuclear Single Quantum Coherence (HSQC) spectroscopy An example of an experiment that combines 1H and 13C spectroscopy is the Heteronuclear Single Quantum Coherence (HSQC). Since proton detection is employed, it is much more sensitive than a carbon detected experiment. The resulting spectrum reveals which carbons and protons are chemically bounded (see Figure 3E for methionine). Hence, it can be used to facilitate spectral assignment. For instance, the assignment of 3-methylglutaconic acid, 3hydroxyisovaleric acid and 3-methylglutaric acid in urine from a patient with 3- 39 Chapter 1.3 methylglutaconic aciduria type I (OMIM: 250950) [47]. Furthermore effectiveness in diagnosing Canavan disease has been suggested [45]. its DATA PROCESSING AND SPECTRAL INTERPRETATION The NMR data are typically processed by Fourier transformation following apodization and zero filling of the free induction decay (FID). The phase is then corrected to obtain absorption line shapes and if necessary, a subsequent baseline correction can be performed. Metabolites can be quantified by fitting their resonances and the internal concentration reference signal with Lorentzian line shapes. Subsequently the concentrations are calculated from the relative integrals of the fitted resonances and the known concentration of the internal standard. For blood and CSF metabolite concentrations are expressed in µmol/l. For urine, the creatinine singlet at 3.13 ppm (3 equivalent protons) is commonly used as internal reference and metabolite concentrations are usually reported in µmol/mmol creatinine. The NMR spectrum of a body fluid taken from a patient with an inborn error of metabolism can be different from a control spectrum in three ways. First, unusual metabolites can be observed. Secondly, certain metabolites that are always present can be totally or partially absent. Thirdly, a novel constituent of the body fluid occurs in increased concentration compared to an age related reference range. For relatively sparse spectral regions, this information can be directly obtained from a 1H 1D spectrum and consequently lead to the diagnosis. However, body fluid spectra are usually very complex (especially for urine), showing hundreds to thousands of resonances in which small biochemical changes may be easily lost. Automatic data reduction and pattern recognition methods reduce this wealth of information and can therefore provide an alternative way for the diagnosis of inherited metabolic diseases [27;48]. The first step is dividing the 1D NMR spectrum into small sections, which provide a series of descriptors for each sample. These descriptors are then used to map the spectrum using non-linear mapping (NLM) or principal component analysis (PCA). The resulting twodimensional map displays clusters of similar samples. A nice separation from control urine samples was obtained for patients suffering from cystinuria, oxalic aciduria, porphyria, Fanconi syndrome and 5-oxoprolinuria [48]. It should be noted that different classification parameters can result in different clustering. Therefore, several maps may be required when screening for distinct metabolic diseases. 40 NMR spectroscopy of body fluids NMR spectra of body fluids HEALTHY INDIVIDUALS AND DATABASES OF MODEL COMPOUNDS Urine Figure 4 shows a part of the 500 MHz 1H-NMR spectra (4.4 – 0.8 ppm) of urine from a healthy child and a healthy adult. In both spectra, the major metabolites and the peaks with chemical shift at pH 2.5 are creatinine (singlet: 3.13 ppm; singlet: 4.29 ppm), creatine (singlet: 3.05 ppm; singlet 4.11 ppm), citric acid ([AB]2 spin system: 2.99; 2.83 ppm), dimethyl amine (triplet: 2.71 ppm), alanine (doublet: 1.51 ppm) and lactic acid (doublet: 1.41 ppm). Figure 4 500 MHz 1H-NMR spectrum of urine from a healthy adult (A) and a healthy child (B). High concentrations of betaine and N,N-dimethylglycine (DMG) can be observed in the urine from the child (B). These concentrations are normal for age. Several sources with peak assignments, chemical shift and multiplicity for a wide range of metabolites are available in literature [5;6;24;49]. For the interpretation of urine NMR spectra the age dependency in the concentration of many metabolites is of relevance. For instance, high concentrations of betaine (singlet: 3.27 ppm; singlet: 3.94 ppm) and N,N-dimethylglycine (DMG) (singlet 2.93 ppm) 41 Chapter 1.3 can be observed in the urine 1D 1H spectrum from the child (Figure 4B). Especially in the first month of life values up to 1500 μmol/mmol and 550 µmol/mmol creatinine may be found for betaine [50] and DMG respectively [42]. In adult urine, the concentration of betaine and DMG generally is lower with a maximal normal value of 80 µmol/mmol creatinine and 26 µmol/mmol creatinine, respectively [42]. Betaine and DMG are metabolites that will not be detected using conventional screening techniques and NMR studies have reported age-related reference values for these metabolites [42;50]. Beside age, other factors such as gender, medication and diet may cause severe variations in body fluid NMR spectra. These factors will be discussed on page 52 (Drugs and food components in body fluids). Plasma/serum In deproteinised blood plasma, the major metabolites and the peaks with chemical shift at pH 2.5 are lactic acid (doublet: 1.41 ppm, quartet: 4.36 ppm), citric acid ([AB]2 spin system: 2.99; 2.83 ppm), glucose (doublet of doublets: 3.23 ppm, various: 3.30 – 3.95 ppm, doublet: 4.63 ppm, doublet: 5.22 ppm), acetic acid (singlet 2.08 ppm) (Figure 2A). Due to the high concentration of glucose (5 mmol/l), identification of metabolites in the region between 3.30 and 3.95 ppm is difficult with 1D 1H NMR spectroscopy. For instance, the singlet resonance of glycine at 3.72 ppm, caused by the 2 protons of the methylene group, is hard to identify at physiological concentration (concentration range: 100 – 384 µmol/l). NMR data of 38 identified metabolites and 14 unidentified peaks measured in deproteinesed blood plasma at pH 2.5 have been published [4]. Recently, the singlet resonance at 3.14 ppm that was described as unknown in this blood plasma study has been identified as dimethyl sulfone (DMSO2) [10]. Several other reports about assignments in human blood plasma have been published [23;51]. Cerebrospinal fluid In literature, there are several 1H-NMR studies of CSF [3;17;19;20;52]. Sweatman et al. [17] reported assignments of 138 resonances in 1H-NMR spectra of human CSF. These derived from 46 metabolites and the assignments were achieved by the measurement of a combination of 2D experiments (JRES and COSY spectra). A standardized method for recording 1D 1H-NMR spectra from CSF was published by Wevers et al. [3]. Quantitative analysis by NMR vs. conventional techniques showed good correlation for alanine, threonine, valine and lactic acid. Resonances from 50 metabolites in 40 CSF were found in this study. Furthermore, 16 unidentified resonances were reported. Myo-inositol (triplet: 3.27 ppm; triplet: 4.05 ppm) in CSF is strongly age-dependent. The concentration of myo-inositol may be as high as 700 µmol/l in the CSF of young children (0 – 3 months) while in adults the upper reference range limit is estimated to be < 30 µmol/l. 42 NMR spectroscopy of body fluids NMR SPECTRA OF KNOWN INBORN ERRORS OF METABOLISM Figure 5 shows an example of a 1D 1H-NMR spectrum of a urine sample from a patient with an IEM. The patient has alkaptonuria (OMIM: 203500). The disease is caused by a deficiency of the enzyme homogentisic acid oxidase in tyrosine metabolism (Figure 6, enzyme H). The characteristic metabolite for this disease is homogentisic acid. Normally this compound cannot be detected in body fluids. The concentration in this patient’s urine amounted to 2500 μmol/mmol creatinine. Figure 5 500 MHz 1H-NMR spectrum of urine from a patient with Figure 6 Metabolic pathway of tyrosine. alcaptonuria. The spectrum shows a high concentration of Homogentisic acid oxidase deficiency detected in urine. homogentisic homogentisic acid (2500 μmol/mmol creatinine) normally not (black square) results in accumulation of acid in patients with alkaptonuria. T: Tyrosine transaminase P: 4-Hydroxyphenylpyruvate dioxygenase H: Homogentisic acid oxidase Table 2 provides a list of available papers on metabolic diseases studied with NMR spectroscopy. In many studies, peak assignments were achieved by the measurement of two-dimensional experiments, measurement of the pure model compound, adding the pure model compound to the body fluid or specific resonance shifts upon pH changes. In the next pages we provide an overview of some key publications on several groups of inherited metabolic disorders studied with NMR spectroscopy. The available literature indicates the power of NMR spectroscopy for the field of inborn errors of metabolism. This review shows that at least 100 different IEM can be diagnosed by NMR spectroscopy of urine. For most of these diseases the handbook of NMR spectroscopy in inborn errors of metabolism shows the characteristic parts of the NMR spectrum [5]. 43 Chapter 1.3 Table 2: Inborn errors of metabolism studied using body fluid NMR spectroscopy. Inborn error of metabolism References Amino acid disorders Aminoacylase deficiency, Type I [5;53-56] Argininosuccinic acid lyase deficiency [5;46;57;58] Citrullinaemia [5;11;57] Cystathionine beta synthase deficiency [5] Cystinuria [48;59;60] Hawkinsinuria [5;61] Histidinaemia [3;5] Homocystinuria [62-64] Hyperlysinaemia II [5] Lysinuric protein intolerance [5] Methionine adenosyl transferase deficiency [5;10] Non-ketotic hyperglycinaemia [5;24] Ornithine carbamoyl transferase deficiency [5;57] Phenylketonuria [5;24;65] Prolidase deficiency [5;66] Prolinaemia II [5;67] Sarcosinaemia [5] Tyrosinaemia I [5;11] Carbohydrate and related disorders Congenital Defect of Glycosylation type IIb [5;68] Galactosaemia [5;33;41] L-Arabinosuria [5;69] Ribose 5-phosphate isomerase deficiency [5;33;70] Lysosomal diseases β- Mannosidosis [5;11] α-Mannosidosis [5;11] Aspartylglucosaminuria [5;11] Fucosidosis [5;11] GM1-gangliosidosis [5;11] GM2-gangliosidosis (Sandhoff Disease) [5;11] Salla disease [5;11;34;71] Sialidosis [5;11] Organic acidurias β-Ketothiolase deficiency [5;24] 2-Aminoadipic aciduria [5] 2-Hydroxyglutaric aciduria [5;24;27;43;72] 2-Methyl-3-hydroxybutyryl CoA dehydrogenase deficiency [5] 2-Methylacetoacetyl CoA thiolase deficiency [38] 3-Hydroxy-3-methylglutaryl CoA lyase deficiency [5;24;32;47;73] 3-Methylcrotonyl CoA carboxylase deficiency, isolated [3;5;24;47;74] 3-Methylglutaconic aciduria, Type I [5;47] 44 NMR spectroscopy of body fluids Table 2: Inborn errors of metabolism studied using body fluid NMR spectroscopy. (Continued) 4-Hydroxybutyric aciduria [5] 5-Oxoprolinuria [4;5;48;72;73;75] Alkaptonuria [5;59;76-78] Canavan disease [3;5;11;44;45;79] Cytochrome c oxidase deficiency [80] Glutaric aciduria type I [5;39] Glutaric aciduria type II [5;24;81] Isovaleric acidaemia [5;24;39;47;82;83] Lactic acidosis [5;24] Malonic aciduria [5] Maple syrup urine disease (=branched-chain ketoaciduria) [5;24;27;39;47;73] Methylmalonic aciduria [5;24;39;82;84-87] Mevalonic aciduria [5] Oxalic aciduria [48;73] Propionic acidaemia [5;24;39;85;86;88] Pyruvate dehydrogenase E1 deficiency [80] Pyruvate dehydrogenase E3 deficiency [80] Lipid metabolism Cerebrotendinous xanthomatosis (CTX) [5;89] Glycerol kinase deficiency [5;59] Medium-chain acyl CoA dehydrogenase (MCAD) deficiency [90] Refsum disease [89] Short-chain acyl CoA dehydrogenase (SCAD) deficiency [5] Sitosterolemia [89] Smith-Lemli-Opitz syndrome (SLOS) [89] Purines / Pyrimidines Disorders Adenosine deaminase deficiency [5;91] Adenylosuccinate lyase deficiency [5;91] β-Ureidopropionase deficiency [5;36;92] Dihydropyrimidinase deficiency [5;91;93;94] Dihydropyrimidine dehydrogenase deficiency [5;91;93;94] Molybdenum cofactor deficiency [5;91] Purine nucleoside phosphorylase deficiency [5;91] UMP synthase deficiency [5] Xanthine dehydrogenase deficiency [5;91] Vitamin related disorders Biotinidase deficiency [5;24] Cobalamin Disorder Cbl-C [95] Vitamin B12 deficiency [20] Others Cytochrome b deficiency [80] Dimethylglycine dehydrogenase deficiency [5;42;96;97] Fanconi syndrome [48] French type sialuria [5;11] 45 Chapter 1.3 Table 2: Inborn errors of metabolism studied using body fluid NMR spectroscopy. (Continued) Guanidinoacetate methyltransferase deficiency [5;98] Huntington's disease [99] Hypoacetylaspartia [5;100] N-Acetylaspartylglutamate accumulation in CSF [5;25] Porphyria [48] Trimethylaminuria (fish-odour syndrome) [5;101-107] Wilson disease [108] Amino acid disorders Amino acids are metabolites containing an amino group (NH2) and a carboxylic acid group (COOH). They are extensively metabolized and several inborn errors of metabolism are located in amino acid metabolism. One of such pathways is the biosynthesis of urea (urea cycle). Investigation of urea cycle disorders by 1H-NMR spectroscopy was published for the urine of patients with citrullinaemia (OMIM: 215700), argininosuccinic aciduria (OMIM: 207900) and ornithine carbamoyl transferase deficiency (OMIM: 311250) [57]. The diagnostic metabolites citrulline, N-acetylcitrulline could be observed in the urine 1H-NMR spectrum of all patients with citrullinaemia. These observations agree with those of Engelke et al. [11] who reported an urinary citrulline and N-acetylcitrulline concentration of 825 and 518 µmol/mmol creatinine, respectively in a patient with argininosuccinate synthetase deficiency (citrullinemia). The urine 1D spectrum of the patients with argininosuccinic aciduria shows the characteristic metabolite argininosuccinate. Orotic acid was detected in urine samples from three out of four patients with ornithine carbamoyl transferase deficiency. For the diagnosis of amino acid disorders cerebrospinal fluid is seldom used. The metabolic profile of CSF may confirm a diagnosis. Table 2 gives an overview of the inborn errors involving amino acid metabolism that have been studied with body fluid NMR spectroscopy. Carbohydrate related disorders Carbohydrates may be found in body fluids in a soluble form as monosaccharides, disaccharides, oligosaccharides and mucopolysaccharides and also as glycolipids. The NMR spectrum of these molecules can be very complex. A paper by Moolenaar [33] gives a table with resonances typical for a variety of monosaccharides and disaccharides. Oligosaccharides accumulate in the urine in several lysosomal diseases. Specific oligosaccharides have been isolated from thin layer chromatography plates or otherwise purified and could subsequently be characterized using NMR spectroscopy. In some cases this has contributed to finding the primary defect in such disease. The most recent example is the identification of a tetrasaccharide Glc(alpha1-2)Glc(alpha1-3)Glc(alpha1-3)Man 46 NMR spectroscopy of body fluids structure that contributed to the identification of the primary defect in the subtype IIb of the CDG syndrome [68] (Congenital Disorder of Glycosylation IIb) (OMIM: 606056). Engelke et al have claimed the diagnosis of at least 5 different lysosomal diseases in native urine with NMR spectroscopy [11]. The ganglioside concentration in the urine of patients with GM1 and GM2 gangliosidosis escaped detection by NMR in the native urine sample probably because of the low concentration. The group of the mucopolysaccharidoses form a subgroup of the lysosomal diseases. The accumulating mucopolysaccharides or glycosaminoglycans are long chains of repeating saccharide units. Hochuli et al. have studied these compounds with twodimensional NMR spectroscopy but direct diagnosis in the native urine is not feasible [109]. Organic acids Organic acids are indicators of organic acidurias associated with various inborn errors of metabolism. In 1985, Iles et al. [39] studied the organic acid metabolites in urine from patients with propionic acidemia, methylmalonic aciduria, branched chain ketoaciduria (maple syrup urine disease), isovaleric acidemia and glutaric aciduria type I. The study provides a list with 1H-NMR chemical shifts (at pH 6.0) of 29 metabolites observed in the urine of the patients with the 5 selected organic acidurias. Another list containing chemical shifts of many organic acid metabolites was published by Lehnert and Hunkler [24]. They measured urine samples and model compounds with NMR spectroscopy under standard pH conditions (pH 2.50). They investigated urine samples of 11 different inborn errors of metabolism involving organic acids including three diseases from leucine metabolism. A recent study gives the body fluid NMR characteristics for all known inborn errors of leucine metabolism [47]. 1H-NMR spectroscopy of urine can easily discriminate between these inborn errors of leucine metabolism and provide a correct the diagnosis. Table 2 gives an overview of the available studies on organic acidurias. Purines and pyrimidines More than 12 enzyme defects in the metabolic pathways of purines and pyrimidines are known to occur [110]. Wevers et al. [91] have studied body fluids from 25 patients with 9 different inborn errors in these pathways. Characteristic abnormalities could be demonstrated in the 1D 1H-NMR spectra of urine samples of all patients. The only exception was case of adenine phosphoribosyl transferase deficiency in which the accumulating metabolite, 2,8-dihydroxyadenine, could not be seen under the NMR conditions used. In 2,8-dihydroxyadenine all protons exchange rapidly with water and therefore this metabolite is NMR invisible. The authors have provided a list of the most important 1H-NMR resonances from relevant purine and pyrimidine metabolites measured at pH 2.5. 47 Chapter 1.3 Others NMR spectroscopy can be used to find diseases characterized by the presence of metabolites that cannot be detected with conventional screenings methods. The available literature shows that NMR spectroscopy is the method of choice to diagnose patients with trimethylaminuria or “fish odor syndrome” (OMIM: 602079). This is a malodor syndrome caused by the accumulation of trimethylamine (TMA), due to a deficiency of the liver enzyme flavin-containing monooxygenase 3 (FMO3). In urine of affected patients the TMA concentration is increased while the trimethylamine N-oxide (TMAO) concentration is decreased. NMR spectroscopy enables the simultaneous determination of these metabolites in urine (TMA: doublet 2.98 ppm and TMAO: singlet at 3.54 ppm). Patients with milder mutations can only be diagnosed with a loading test as the TMA concentration in a random urine sample may be fully normal. Loading tests have been described with TMA [111], with choline [112] and with fish meal [104]). Guanidinoacetate methyltransferase (GAMT) deficiency (OMIM: 601240) deficiency is a disorder in the biosynthesis of creatine. Guanidinoacetic acid (GAA) is the characteristic accumulating metabolite in GAMT deficiency, the oldest of the three creatine biosynthesis disorders known to date. 1H-NMR spectroscopy of urine and CSF can be used for diagnosing this disorder [5;98]. The NMR spectrum of the urine of a patient shows elevated GAA (doublet at 3.98 ppm). In CSF, creatine (singlet resonances at 3.05 and 4.11ppm) and creatinine (singlet resonances at 3.13 and 4.29 ppm) are completely absent. Normally, they are always present in CSF (reference value: creatine 25 – 70 µmol/l and creatinine 20 - 100 μmol/l [98]). Accumulation of GAA in GAMT patients (CSF: ± 1 µmol/l; plasma: ± 10 µmol/l [113]) is not visible in plasma and CSF 1H-NMR spectra due to overlap of other resonances and the detection limit of the technique. NOVEL DISEASES IDENTIFIED WITH BODY FLUID NMR Three novel diseases were discovered by body fluid NMR spectroscopy, e.g., dimethylglycine dehydrogenase deficiency (OMIM: 605850), β-ureidopropionase deficiency (OMIM: 606673) and a severe hypomyelination associated with increased levels of N-acetylaspartylglutamate in CSF. In two other cases the technique was strongly involved in the elucidation of the molecular defect, e.g., ribose 5-phosphate isomerase deficiency (OMIM: 608611) and congenital disorder of glycosylation type IIb (OMIM: 606056). The NMR findings in these novel diseases are shortly described below. 48 NMR spectroscopy of body fluids Dimethylglycine dehydrogenase deficiency (OMIM: 605850) In urine and plasma NMR spectra of an adult patient with a fish odor and muscle weakness, high singlet resonances were observed at 2.93 and 3.80 ppm. 1H-NMR spectroscopy of model compounds revealed that N,N-dimethylglycine (DMG) caused these resonances. Addition of pure DMG to the urine sample and 13C-NMR spectroscopy of the patient’s urine and an authentic reference solution of DMG confirmed that the two singlets were caused by DMG. The concentration of DMG amounted to 457 µmol/mmol creatinine in urine (reference range: 1 – 26) and 221 µmol/l in serum (reference range: 1 - 5). The high concentration of DMG was caused by a deficiency of the enzyme dimethylglycine dehydrogenase (DMGDH) converting DMG to sarcosine. A homozygous missense mutation was found in the DMGDH gene of the patient [96]. β-Ureidopropionase deficiency (OMIM: 606673) A new defect in pyrimidine metabolism found with in vitro 1H-NMR spectroscopy was reported by Assmann et al. [92]. Characteristic resonances of 3ureidopropionic acid (triplet: 2.57 ppm; triplet: 3.36 ppm) and 3-ureidoisobutyric acid (doublet: 1.13 ppm; multiplet: 2.69 ppm; AB protons: 3.26 nd 3.29 ppm) were observed in high concentrations in the urine spectrum (861 and 718 μmol/mmol creatinine, respectively) [36]. One-dimensional and two-dimensional NMR techniques were used to identify these metabolites. The accumulation of these metabolites was caused by a deficiency of the enzyme β-ureidopropionase (UP) converting 3-ureidoisobutyric acid and 3-ureidopropionic acid to, respectively, β-aminoisobutyric acid and β-alanine. In this patient, Vreken et al. [114] found compound heterozygosity for 2 splice acceptor site mutations in the UPB1 gene, IVS1-2A-G and IVS8-1G-A. Van Kuilenburg et al. described a group of patients with this defect [115]. Severe hypomyelination associated with increased levels of N-acetylaspartylglutamate in CSF Wolf et al. [25] described two unrelated girls with a severe hypomyelination of the brain. Defects in the proteolipid protein (PLP) gene causing X-linked PelizaeusMerzbacher disease (PMD) could be ruled out. NMR spectroscopy of the CSF revealed an increased concentration of the dipeptide N-acetylaspartylglutamic acid (NAAG), a putative neurotransmitter that is thought to play a role in intercellular signalling in the brain. NAAG concentration amounted to 95 and 197 µmol/l in the 2 patients (reference: < 4.4 µmol/l). The molecular and genetic basis of this novel disease remains as yet unsolved. This is the first inborn error of metabolism where a first key for understanding the molecular defect was found with NMR spectroscopy of CSF. 49 Chapter 1.3 Ribose 5-phosphate isomerase deficiency (OMIM: 608611) The combination of in vivo NMR spectroscopy of the brain with body fluid in vitro NMR spectroscopy led to the recognition and characterization of the first ever inborn error of polyol metabolism. In vivo NMR spectroscopy of the brain with a severe white matter disease showed unknown resonances between 3.5 and 4.0 ppm. Body fluid NMR spectroscopy identified increased amounts of arabitol and ribitol in all body fluids of the patient. Arabitol and ribitol in urine amounted to 1800 and 190 µmol/mmol creatinine (reference respectively: <98 and <11). In CSF arabitol and ribitol concentrations were 5100 and 1200 µmol/l (reference range respectively: <39 and <5 [33]). It could be shown that polyols caused the abnormal resonances in the brain MR spectrum. This led to the discovery of the defective enzyme ribose 5-phosphate isomerase, an enzyme in the pentose phosphate pathway. Mutations in the corresponding gene could be demonstrated in the DNA [116]. Congenital Disorder of Glycosylation type IIb (OMIM: 606056) The group of van Coster found a patient who constantly excreted an unknown tetrasaccharide in the urine. This compound could be shown with thin-layer chromatography. After isolation NMR spectroscopy helped to identify this compound as Glc(α1-2)Glc(α1-3)Glc(α1-3)Man. This led the way to the characterization of a novel subtype of the so called CDG syndrome (Congenital Disorder of Glycosylation type IIb). The defect is in the enzyme glucosidase I (GCS1) involved in biosynthesis of N-glycans. The defect could be confirmed at the enzyme level and the molecular genetic level [68]. Abnormal resonances from the accumulating tetrasaccharide can be directly observed in a random sample of the patient [5]. NOVEL PATHWAY: METHANETHIOL METABOLISM Using one-dimensional and heteronuclear two-dimensional NMR techniques, Engelke et al. [10] have shown that dimethyl sulfone (DMSO2) is a "regular" constituent of CSF (Figure 7) and plasma. DMSO2 derives from dietary sources, intestinal bacterial metabolism, and from human endogenous metabolism. In man, methanethiol is converted into DMSO2. Increased plasma DMSO2 concentrations are found in patients with methionine adenosyltransferase I/III (MAT I/III) deficiency (OMIM: 250850). These findings led to the elucidation of a novel metabolic pathway that was not known to occur in man (Figure 8). 50 of dimethyl sulfone (singlet 3.14 ppm) amounted to 18 µmol/L. peak with the coordinates 3.14 (1H) / 44.3 (13C). The concentration cerebrospinal fluid sample (pH=2.5). Dimethyl sulfone showed a 1D1H (upper) and 1H-13C HSQC (lower) spectrum of a representative Figure 7 Both, dietary sources and putative endogenous metabolic pathways can potentially contribute the presence of dimethyl sulfone in human body fluids. The partial putative pathway leading to dimethyl sulfone in human body fluids. Figure 8 Chapter 1.3 DRUGS AND FOOD COMPONENTS IN BODY FLUIDS Beside age, other factors such as medication and diet may cause severe variations in body fluid NMR spectra and hinder proper spectrum interpretation and quantification. Some examples are described below. Drugs Valproic acid (Depakene®, Valproate, Valrelease®). Several NMR studies [117-119] reported urine samples from patients who took the anticonvulsant drug valproic acid. In man, valproic acid is extensively metabolized into several products, and some are excreted into the urine. The urine 1H-NMR spectra of these patients show many high resonances from 3-hydroxyvalproic acid and valproic acid glucuronide. Azaroual et al. [118] reported a valproic acid intoxification identified by NMR spectroscopy. Vigabatrin (Sabril®) The anti-epileptic drug vigabatrin is often used in patients suspected to have an inborn error of metabolism. The 1H-NMR urine spectra of vigabatrin treated patients show high vigabatrin resonances between 2.0 and 2.1 ppm caused by CH2 protons (multiplet). This vigabatrin multiplet interferes with the N-acetyl resonances. Therefore correct interpretation and quantification of the N-acetyl region of the 1H-NMR spectrum is impossible. Ethosuximide (Zaronthin®) Ethosuximide is a drug used to control seizures. A patient with guanidinoacetate methyltransferase deficiency was treated with this drug and resonances of ethosuximide (triplet: 0.86 ppm; singlet: 1.28 ppm; multiplet: 1.65 and AB: 2.68 ppm) could be observed in the CSF 1H-NMR spectrum of this patients [98]. Due to the enzyme deficiency, creatine and creatinine resonances were absent in this spectrum as expected. Dimethyl sulfone (Methylsulfonylmethane, MSM) NMR spectroscopy was performed on CSF from a 3-year-old boy with mental retardation and loss of acquired skills. A high concentration of DMSO2 (singlet: 3.14 ppm) was observed in his 1D 1H-NMR CSF spectrum. At the time of the lumbar puncture the patient had used a dietary supplement containing DMSO2 [10]. Other drugs that are observed in body fluid NMR spectra are acetaminophen (Paracetamol®) [14;119-121], ampicillin [122], carbamazepine (Epitol®, Tegretol®) [119], diethylcarbamazine (Hetrazan®) [123] and penicillin [14] and the drug vehicle, propylene glycol [52]. In some cases also metabolites of the drugs can be observed. 52 NMR spectroscopy of body fluids Figure 9 500 MHz 1D 1H-NMR spectrum of urine from a healthy control after drinking Coca Cola light. Acesulfame K showed a doublet at 2.11 ppm and a quartet at 5.67 ppm. Food components Acesulfame K Acesulfame K is a sweetener about 200 times sweeter than sucrose. Acesulfame K is not metabolized and it can be found in the urine 1H-NMR spectra of persons who drink light products containing Acesulfame K, e.g., cola light (Figure 9). Other food ingredients or their metabolites that are observed in body fluid NMR spectra are ethanol in alcoholic drinks (triplet: 1.17 ppm; quartet: 3.64 ppm), TMA Noxide in fresh see fish (singlet: 3.54 ppm), trigonelline in coffee (singlet: 4.44 ppm; triplet: 8.10 ppm; singlet: 9.19 ppm) and Vitamin C (multiplet: 3.75 ppm; multiplet: 4.07 ppm; doublet: 4.97 ppm). 53 54 TWO N-ACETYLATED METABOLITES IN BODY FLUIDS 55 56 CHAPTER 2.1 N-Acetylated Metabolites in Urine: Proton Nuclear Magnetic Resonance Spectroscopic Study on Patients with Inborn Errors of Metabolism Udo F.H. Engelke, Maria L.F. Liebrand-van Sambeek, Jan G.N. de Jong, Jules G. Leroy, Éva Morava, Jan A.M. Smeitink and Ron A. Wevers Clinical Chemistry 2004; 50(1): 58-66 57 58 N-acetylated metabolites in urine Abstract Background: There is no comprehensive analytical technique to analyze Nacetylated metabolites in urine. Many of these compounds are involved in inborn errors of metabolism. In the present study, we examined the potential of proton nuclear magnetic resonance (1H-NMR) spectroscopy as a tool to identify and quantify N-acetylated metabolites in urine of patients with various inborn errors of metabolism. Methods: We performed 1H-NMR spectroscopy on a 500 MHz spectrometer. Using a combination of one- and two-dimensional correlation spectroscopy (COSY) 1HNMR spectra, we were able to assign and quantify resonances of characteristic Nacetylated compounds products in urine of patients with 13 inborn errors of metabolism. Results: The disease-specific N-acetylated metabolites were excreted at concentrations >100 µmol/mmol of creatinine in the patients’ urine. In control urine samples, the concentration of individual N-acetyl-containing compounds was <40 µmol/mmol of creatinine. The combination of one- and two-dimensional COSY NMR spectroscopy led to the correct diagnosis of nine different inborn errors of metabolism. No abnormalities were observed in the spectra of urine from patients with GM1- or GM2-gangliosidosis. We also determined the 1H-NMR characteristics of N-acetylated metabolites that may be relevant to human metabolism. Conclusion: spectroscopy may be used to identify and quantify Nacetylated metabolites of diagnostic importance for the field of inborn errors of metabolism. 1H-NMR 59 Chapter 2.1 Introduction Proton nuclear magnetic resonance (1H-NMR) spectroscopy of body fluids has been used to diagnose inborn errors of metabolism [3;4;39]. The technique yields an overview of proton-containing metabolites, is rapid and nondestructive, and requires minimal sample pretreatment. In this report, we document the effective use of 1H-NMR spectroscopy for the identification and quantification of N-acetylcontaining metabolites in body fluids of patients with various inborn errors of metabolism. N-Acetylation occurs in many metabolic pathways, which makes the detection of N-acetylated compounds important. Several diseases are known to be associated with the accumulation of N-acetylated metabolites. Some of these have been studied by 1H-NMR spectroscopy. The urinary excretion of N-acetylated amino acids has been published for patients with inborn errors of amino acid metabolism [124;125]. N-Acetylaspartic acid and N-acetylaspartylglutamic acid are the laboratory marker metabolites in Canavan disease [126]. Other authors have described one- and two-dimensional NMR experiments for structural characterization and identification of N-acetylneuraminic acid-containing storage products in lysosomal diseases [127-129]. These studies demonstrate that the detection of N-acetylated metabolites assists in the diagnosis of various inborn errors of metabolism. However, many N-acetylated compounds remain undetected by the conventional analytical techniques used in metabolic screening laboratories. The three equivalent N-acetyl protons resonate as a singlet in a small part of the 1H-NMR spectrum (1.9–2.2 ppm). Only a few other metabolites are known to cause resonances in this area. The finding of singlet signals in a one-dimensional proton NMR spectrum can thus be used as an indication of the presence of an Nacetylated metabolite. We present here the NMR characteristics of many N-acetylated compounds relevant for human metabolism. In addition, we provide NMR spectra for urines representative for 13 inborn errors of metabolism that involve N-acetylated compounds. Materials and Methods MODEL COMPOUNDS Most model compounds were commercially available from Sigma or Aldrich. NAcetylated oligosaccharides from lysosomal storage diseases were isolated preparatively from thin-layer chromatography (TLC) plates routinely used for mono- and oligosaccharide analysis from urine of affected patients. Manα13Manβ1-4-N-acetylglucosamine (GlcNAc), Fuc-GlcNAc-Asn, aspartylglucosamine, and two sialyloligosaccharides were isolated with a butanol–acetic acid–water (2:1:1 60 N-acetylated metabolites in urine by volume) solvent system and visualized with orcinol–sulfuric acid (Fig. 1). Because of an overlap with lactose, Manβ1- 4GlcNAc from the urine of a βmannosidosis patient was isolated with propanol–acetic acid–water (42.5:0.5:7.5 by volume) solvent system. The characteristic TLC bands were removed preparatively from the silica-coated plate and measured as a model compound. Figure 1 Thin-layer chromatogram of dextran and urine samples. The solvent system was butanol–acetic acid–water and bands were visualized using orcinol-sulfuric acid staining. Lanes: 1, dextran; 2, sample from a patient with α-mannosidosis; 3, sample from a patient with fucosidosis; 4, sample from a patient with GM2-gangliosidosis; 5, sample from a patient with GM1-gangliosidosis; 6, dextran; 7, sample from a healthy individual; 8, sample from a patient with aspartylglucosaminuria; 9, sample from a patient with sialidosis. The arrows indicate the characteristic TLC bands. URINE Control urine samples were from 10 healthy individuals and from 50 diseased patients. In the urine samples from diseased controls sent to our laboratory, extensive metabolic screening had ruled out any relevant inborn errors of metabolism. Urine was centrifuged before analysis. We added 70 µL of a 20.2 mmol/l trimethylsilyl-2,2,3,3-tetradeuteropropionic acid (sodium salt; Aldrich) 2H2O solution to 700 μL of urine as a chemical shift reference (δ = 0.00) and as a lock signal. The pH of the urine was adjusted to 2.50 ± 0.05 with concentrated HCl. Finally, we placed 650 µL of the sample in a 5-mm NMR tube (Wilmad Royal Imperial). 61 Chapter 2.1 NMR MEASUREMENTS High-resolution spectra for urine samples and the model compounds were acquired at 500 MHz on a Bruker DRX and AMX spectrometer equipped with a sample changer. The spin-lattice or longitudinal relaxation time (T1) values for Nacetylneuraminic acid, N-aspartylglucosamine, N-acetylaspartic acid, and Man(14)GlcNAc, as representative N-acetylated compounds, were determined in urine of patients by the inversion-recovery technique. The T1 values amounted to 1.0, 0.9, 1.6, and 0.9 s, respectively. ONE-DIMENSIONAL 1H-NMR SPECTROSCOPY For urine, 128 transients were collected into 32 000 data points with a spectral width of 6002 Hz. The H2O resonance was presaturated during the relaxation delay (10 s). Shimming of the sample was performed automatically on the deuterium signal. A sine-bell squared filter (SSB = 2) was used, and spectra were Fourier-transformed after the free induction decay was zero-filled to 64 000 data points. The phase and the baseline were corrected manually. The N-acetyl resonances in the one-dimensional urine spectra were fitted semiautomatically to a Lorentzian line shape model function. The integrals of these fits were used for quantification by comparing them to the fit integral of creatinine (singlet at 3.13 ppm). For identification and quantification of the NMR spectra, 1D WinNMR and the AMIX software were used (Bruker BioSpin). A prerequisite for accurate quantitative determination of metabolites with proton NMR spectroscopy lies in the length of the relaxation delay. We used a delay of 10 s, which is more than five times the T1 value of the N-acetyl protons in Nacetylaspartic acid, the compound with the longest T1 value among the Nacetylated compounds. 1H-1H CORRELATION SPECTROSCOPY (COSY) The spectral widths in the F1 and F2 axes were 6002 Hz, and 4000 data points were collected in F2. For the urine samples, 256 increments and 16 scans per increment were used. The relaxation delay was 2 s, and before the Fourier transformation, a sine function was applied in both time domains. During the relaxation delay, the water resonance was presaturated. QUANTITATIVE DATA To study the linearity and recovery, we added three N-acetylated metabolites, N-a cetyltyrosine (cat. no. A-2513; Sigma), N-acetyltryptophan (cat. no. A-6376; Sigma), and N-acetylaspartic acid (cat. no. A-5625; Sigma), separately to normal human urine in concentrations of 0, 100, 500, and 3000 µmol/l. After addition of these metabolites the creatinine concentration of the urine was 1.1 mmol/l. The same experiment was repeated with a second urine with a creatinine concentration of 2.0 mmol/l. For concentration calculations for the N-acetylated metabolites 62 N-acetylated metabolites in urine using 1H-NMR spectroscopy, we used trimethylsilyl-2,2,3,3-tetradeuteropropionic acid as the concentration reference, and the metabolite concentrations were expressed in µmol/l. Results N-ACETYL RESONANCES IN URINE A representative 1H-NMR spectrum of urine from a healthy individual is shown in Figure 2. Methyl protons from N-acetyl groups characteristically give a singlet resonance in a limited part of the 1H-NMR spectrum (1.9–2.2 ppm). Figure 2 500 MHz 1H-NMR spectrum of urine from a healthy volunteer. The creatinine peak (singlet at 3.13 ppm) is scaled to 100. The inset show the N-acetyl region in greater detail. Table 1 provides an overview of the N-acetyl-containing metabolites detectable in urine samples from healthy individuals and patients. Individual compounds can be recognized by the chemical shift of the N-acetyl group CH3 protons and from resonances deriving from other parts of the molecule. Neuraminic acid, Nacetylaspartic acid, and the N-acetyl groups in glycoproteins in urine from healthy controls are known to contribute to the resonances in this area [6]. In addition, a few metabolites without an N-acetyl group are known to resonate between 1.9 and 2.2 ppm. These include glutamine (β-CH2 multiplet at 2.15 ppm), proline (β-CH2 multiplet at 2.02 ppm), methionine (S-CH3 singlet at 2.13 ppm) [130], and acetic acid (CH3 singlet at 2.08 ppm). QUANTITATIVE DATA The linearity was excellent for all three N-acetylated metabolites, N-acetyltyrosine, N-acetyltryptophan, and N-acetylaspartic acid (slope, 0.94–1.09; intercept did not 63 Chapter 2.1 significantly deviate from zero; correlation coefficient >0.99). The recovery was in 90–110% for these compounds at all concentrations. To study the reproducibility (within-run), we prepared and measured 10 samples from the same urine on the same day. The within-run CV at 3000 µmol/l was <4% for all three N-acetylated metabolites. Table 1: 1H Resonances positions and multiplicity of N-acetyl containing metabolites. Metabolites Resonance position (ppm) and multiplicity Acetyl CH3 singlet (s) Rest of the molecule N-Acetylneuraminic acid (β anomer) 2.05 1.85 (dd); 2.26 (dd) N-Acetylneuraminic acid (α anomer) 2.03 1.65 (dd); 2.72 (dd) 2.02, 2.03, 2.06 1.72 (dd); 2.67 (dd) N-Acetylneuraminic acid (Bound AcNeu2→6) a GlcNAc-Asn 2.01 2.97 (mu); 3.83 (mu); 4.11 (tr); 5.08 (do) (GalGlcNAc-)Asn 2.00 2.95 (mu); 3.54 (mu); 3.87 (mu); 4.10 (tr); 4.48 (do); 5.09 (do) Manβ1-4GlcNAc 2.03 3.72 (mu); 3.89 (mu); 4.72 (do); 5.20 (do) N-Acetylcitrulline* 2.03 1.52 (mu); 1.65-1.80 (mu); 3.12 (qu) N-Acetylaspartic acid 2.03 2.95 (mu); 4.72 (mu) N-Acetyltyrosine 1.94 3.07 (AB); 7.89 (AB) Manα1-6Manβ1-4GlcNAc 2.03 3.88 (mu); 4.04 (do); 5.10 (mu); 5.20 (do) Fuc-GlcNAc-Asn 2.01 1.21 (do); 2.94 (mu); 3.78 (do); 4.06 (tr); 4.12 (qu); 4.90 (do); 5.10 (do) 2.05, 205 β-GlcNAc(1-2)αMan(1-3)\ 4.10 (mu); 4.19 (mu); 4.24 (mu); 4.53 (do); 5.11 (mu); 5.20 (do) βMan(1-4)GlcNAc β-GlcNAc(1-2)αMan(1-3)/ β-GlcNAc(1-2)αMan(1-3)\ 2.04, 2.05, 2.06 βGlcNAc(1-4) - βMan(1-4)GlcNAc 4.10 (mu); 4.19 (mu); 4.24 (mu); 4.45 (do); 4.53 (do); 5.05 (mu); 5.11 (mu); 5.20 (do) β-GlcNAc(1-2)αMan(1-3)/ GalNAcGalFuc (blood group A trisaccharide) 2.04 1.21 (do); 1.23 (do); 4.42 (mu); 5.13 (do); 5.15 (do); 5.18 (do); 5.29 (do); 5.38 (do); 5.45 (si) Acetylcarnitine 2.14 3.19 (si) N-Acetylglutamic acid 2.03 2.00 (mu); 2.19 (mu); 2.49 (tr); 4.38 (mu) Bi-antennary octasaccharide 2.04, 2.05, 2.05 (in GM1-gangliosidosis urine) N-Acetylaspartylglutamic acid 4.11 (do); 4.19 (do); 4.25 (do); 4.46 (do); 4.57 (do); 5.11 (mu); 5.20 (do) 2.04 2.01 (mu); 2.23 (mu); 2.48 (mu); 2.85 (AB); 4.44 (mu); 4.72 (mu) N-Acetyltryptophan 1.92 3.31 (m); 4.70 (qu); 7.2-7.7 (varius) a = si, singlet; do, doublet; tr, triplet; qu, quartet; mu, multiplet; dd, doublet doublet; AB, AB-system. STRATEGY FOR QUANTITATIVE INTERPRETATION AND ASSIGNMENT Because of the low concentrations of N-acetylated metabolites, assignment of their singlet resonances in urine from healthy volunteers is generally not possible. 64 N-acetylated metabolites in urine Metabolic diseases that involve N-acetylated compounds may be diagnosed if the concentrations of the characteristic N-acetylated metabolites are high enough to qualify them as abnormal and to allow their identification. Identification requires the spectral information from the protons in the rest of the molecule (Table 1). As a systematic approach, we quantified the highest singlet resonance between 1.9 and 2.2 ppm. Excluded from this quantification was the singlet for acetic acid (mean chemical shift, 2.083 ppm; minimum, 2.081 ppm; maximum, 2.085 ppm; n = 10). It was assumed that CH3 protons from N-acetylated metabolites cause the resonances in this region. Figure 3 Quantification of the N-acetylated compound with the highest concentration in the urine as function of age. |, controls; z, urine of affected patients. 1, Salla disease; 2, French-type sialuria; 3, α-mannosidosis; 4, β-mannosidosis; 5, fucosidosis; 6, aspartylglucosaminuria; 7, GM1-gangliosidosis; 8, GM2-gangliosidosis; 9, sialidosis; 10, Canavan disease; 11, citrullinemia; 12, tyrosinemia type I; 13, tyrosinemia type II. Figure 3 shows the concentrations of individual N-acetyl-containing compounds in urine samples from 50 diseased controls and 10 healthy volunteers. The highest concentration of an N-acetylated compound was generally <40 µmol/mmol of creatinine. Higher concentrations were observed in children under 5 years. Especially in the first month of life, values up to 200 µmol/mmol of creatinine could be found. Figure 3 also shows data on urine samples from patients with 13 different inborn errors of metabolism involving N-acetylated compounds. Below 40 µmol/mmol of creatinine, the CH3 protons from the N-acetyl group are the only resonances from most N-acetylated compounds that can be observed in the spectra. Therefore, assignments based solely on the N-acetyl singlet were not 65 Chapter 2.1 possible. As shown in Figure 3, most patients with inborn errors of metabolism excreted the N-acetylated metabolites typical for their diseases at concentrations >100 µmol/mmol of creatinine. At such concentrations, the typical N-acetyl compound could be assigned with greater certainty because the other protons of the molecule contribute to the assignment. Two-dimensional COSY spectra may help with further identification of relevant metabolites. As an example, Figure 4 shows how specific cross-peaks in a two-dimensional COSY spectrum of the model compound Fuc-GlcNAc-Asn (Figure 4A) can help to identify this metabolite in the urine of a patient with fucosidosis (Figure 4B). The cross-peaks derive from the GlcNAc protons (3.83–5.10 and 3.78–4.90 ppm), the Asn protons (2.94–4.06 ppm), and the Fuc protons (1.21–4.12 ppm). A Figure 4 COSY spectra of model compound Fuc-GlcNAc-Asn obtained from TLC chromatography (A) and urine B from a patient with fucosidosis (B). The characteristic pattern of the model compound Fuc-GlcNAc-Asn is shown in the circled cross-peaks. 66 N-acetylated metabolites in urine INBORN ERRORS OF METABOLISM Infantile sialic acid storage disorder (ISSD; OMIM 269920) and Salla disease (OMIM 604369). ISSD and Salla disease are both characterized by increased concentrations of free sialic acid (N-acetylneuraminic acid) in the urine and by its storage in the lysosomes of different tissues. The molecular basis of these diseases is a defect in the lysosomal sialic acid transporter. Severe mutations in the gene lead to ISSD, whereas milder mutations clinically present as Salla disease. Characteristically, a 20- to 200-fold increase in N-acetylneuraminic acid excretion in urine has been described in infantile cases of the disease. Bound sialic acid typically is within the range obtained for controls. One- and two-dimensional COSY 1H-NMR spectra were obtained from urine of a 16-year-old patient with Salla disease. The onedimensional urine spectrum showed the presence of N-acetylneuraminic acid [singlet at 2.05 ppm; 121 µmol/mmol of creatinine; reference values, 20.2 (11.6) µmol/mmol of creatinine for ages >2 years [131]; Figure 5A]. The COSY spectrum confirmed the presence of free N-acetylneuraminic acid by showing cross-peaks from the pyranose ring protons of the carbon- 3 atom (1.85 and 2.26 ppm; data not shown) [34]. French-type sialuria (OMIM 269921). French-type sialuria is a metabolic disorder caused by a defect the enzyme uridine diphosphate-N-acetylglucosamine 2-epimerase (EC 5.1.3.14), which is involved in N-acetylneuraminic acid synthesis. The defect in the enzyme leads to failing feedback inhibition by N-acetylneuraminic. Free N-acetylneuraminic acid is massively increased in the body fluids of diseased patients. We investigated the urine of a 6-year-old child with this disease. N-Acetylneuraminic acid excretion (Figure 5B) was much higher than in the patient with Salla disease [4275 µmol/mmol of creatinine; reference values, 20.2 (11.6) µmol/mmol of creatinine for ages >2 years [131]]. Because of the very high concentration of Nacetylneuraminic acid, even the pyranose ring protons (3.5–4.1 ppm) and the nitrogen proton (doublet at 8.09 ppm) were clearly observed in the one- and twodimensional spectra (data not shown). α-Mannosidosis (OMIM 248500). Patients with a deficiency of the lysosomal enzyme α -mannosidase (EC 3.2.1.24) have high concentrations of MannGlcNAc (n ≥ 2) oligosaccharides in the cells and urine. Oligosaccharide TLC of the urine from an affected patient showed many mannose-rich oligosaccharide bands. The major oligosaccharide band (Figure 1, lane 2) typical for this defect was isolated preparatively from the silica-coated plate and measured as a model compound by 1H-NMR. The compound was identified by comparison with reference spectra as the trisaccharide Manα13Manß1-4GlcNAc [132]. One- and two-dimensional COSY 1H-NMR spectra were obtained from urine of the patient. In the COSY spectrum, Manα1-3Manß1- 67 Chapter 2.1 4GlcNAc was recognizable as two cross-peaks (4.04–5.10 and 3.88–5.20 ppm), and in the one-dimensional spectrum was recognizable as a singlet at 2.03 ppm (Figure 5C) and doublet at 5.20 ppm. The urinary concentration of Manα13Manß1-4GlcNAc in our patient was 153 µmol/mmol of creatinine. β-Mannosidosis (OMIM 248510). The lysosomal enzyme β-mannosidase is involved in glycoprotein catabolism. A deficiency of this enzyme leads to excessive urinary excretion of the disaccharide Manß1-4GlcNAc. We measured urine samples from two brothers (ages 8 and 9 years) with ß-mannosidosis. Oligosaccharide TLC was performed on urine samples from the boys. The Manß1-4GlcNAc band (data not shown) was isolated preparatively from the silica-coated plate (100 µg) and measured as a model compound by 1H-NMR. The spectrum showed the N-acetyl singlet (2.03 ppm), the 1α and 1β GlcNAc protons (doublets at 4.72 and 5.20 ppm, respectively), and many resonances in the 3.4–4.1 region deriving from the GlcNAc and Man parts of the disaccharide. The one-dimensional NMR spectra of their urine showed the same singlets at 2.04 ppm (Figure 5D) and doublets at 5.21 ppm. The doublet at 4.72 ppm was not visible because of overlap of the water resonance. The COSY spectra from the urine showed characteristic cross-peaks from the 1α and 1ß GlcNAc protons (3.89–5.20 and 3.72–4.72 ppm, respectively; data not shown). The Manß1-4GlcNAc concentrations in the urine of the brothers were 255 and 273 µmol/mmol of creatinine. Another, smaller singlet at 2.03 ppm was also observed in the one-dimensional NMR spectra of the two brothers (56 and 122 µmol/mmol of creatinine). Presumably this compound is sialyl-α2-6-mannosyl-ß1-4-Nacetylglucosamine ( 10% of the concentration of Manß1-4GlcNAc). Van Pelt et al. have identified this trisaccharide in concentrated urine fractions of affected patients [127]. Fucosidosis (OMIM 230000). Because of a deficiency of lysosomal α-L-fucosidase (EC 3.2.1.51), patients with fucosidosis accumulate fucose-containing glycolipids, oligosaccharides, or glycoasparagines in the urine and tissue. Three characteristic oligosaccharides were isolated from urine of an affected patient by TLC (Figure 1, lane 3). These bands were removed from the TLC plate and dissolved in 2H2O. The fractions were analyzed by 1H-NMR spectroscopy and compared with the literature [133]. The one- and two-dimensional spectra of fraction 1 indicated the presence of FucGlcNAc-Asn (Table 1). Because of the low concentrations in the isolated fractions 2 and 3, interpretation was impossible. The characteristic COSY pattern of FucGlcNAc-Asn (cross-peaks at 1.21–4.12 and 2.94–4.06 ppm) was observed in the spectrum for urine from an affected adult patient (Figure 4B). The urinary FucGlcNAc-Asn concentration was 109 µmol/mmol of creatinine (singlet at 2.01 ppm; Figure 5E). 68 N-acetylated metabolites in urine Figure 5 500 MHz 1H-NMR spectra of urine samples. (A), Salla disease; (B), French-type sialuria; (C), α-mannosidosis; (D), β-mannosidosis; (E), fucosidosis; (F), aspartylglucosaminuria; (G), GM1-gangliosidosis; (H), GM2-gangliosidosis; (I), sialidosis; (J), Canavan disease; (K), citrullinemia; (L), tyrosinemia type I. The creatinine peak (singlet at 3.13 ppm; not shown) in the urine spectra is scaled to 100. 69 Chapter 2.1 Aspartylglucosaminuria (OMIM 208400). Aspartylglucosaminidase deficiency or aspartylglucosaminuria is an inborn error of metabolism caused by the deficiency of the lysosomal enzyme N-aspartyl-ßglucosylaminidase (EC 3.5.1.26). This enzyme cleaves the N-acetylglucosamine– asparagine linkage found in a variety of oligosaccharides and glycoproteins. Patients with the disease excrete aspartylglucosamine (GlcNAc-Asn) as major abnormal metabolite. In addition, several other glycoasparagines, such as Galß14GlcNAcß1-N-Asn, the mannose-rich derivative Man-Man-GlcNAc-GlcNAc-Asn, and NeuAc-Gal-GlcNAc-Asn, can be excreted [134]. Urine samples from two unrelated cases were available. TLC with orcinol–sulfuric acid showed the abnormal Galß1-4GlcNAcß1-N-Asn band in these samples (Figure 1, lane 8). TLC plates stained with ninhydrin showed the GlcNAc-Asn and the Galß1-4GlcNAcß1N-Asn band in the urine (data not shown). The NMR spectra of nearly pure compounds were compared with reference spectra from the literature [135]. The resonances of both compounds are shown in Table 1. The characteristic crosspeaks of both compounds were observed in COSY spectra of urine from both patients. The GlcNAc-Asn concentrations (singlet at 2.02 ppm; Figure 5F) were 169 and 323 µmol/mmol of creatinine, respectively. The Galß1-4GlcNAcß1-N-Asn (singlet at 2.01 ppm) concentrations were 27 (case 1) and 114 µmol/mmol of creatinine (case 2). GM1-Gangliosidosis (OMIM 230500). A deficiency of β-D-galactosidase (EC 3.2.1.23) leads to GM1-gangliosidosis. Oligosaccharides with a D-galactosyl group at the nonreducing end are increased in the urine of affected patients. Three bands were isolated preparatively by TLC from the urine of an affected patient (Figure 1, lane 5) and compared with spectra from the literature [135]. The NMR spectrum of band 2 corresponded to the biantennary octasaccharide [135], and the NMR spectrum of band 3 corresponded to the triantennary decasaccharide [135] or the tetraantennary dodecasaccharide [135]. For band 1, we could not find an exact match with reference spectra. The one-dimensional NMR spectrum of urine from an affected patient (Figure 5G) showed singlets between 2.00 and 2.04 ppm. The singlet with the highest intensity (2.03 ppm) represented a N-acetylated storage compound with a concentration of 34 µmol/mmol of creatinine. GM2-Gangliosidosis (Sandhoff disease; OMIM 268800). As a result of a deficiency in β-N-acetylhexosaminidase (EC 3.2.1.52), oligosaccharides containing N-acetylglucosamine are excreted in the urine of affected patients. Using TLC, we were able to preparatively isolate two bands from the urine of an affected patient (Figure 1, lane 4). The NMR spectra of both isolated bands showed resonances at 5.20, 5.11, and 4.53 and a singlet at 2.05 ppm. Band 2 differed from band 1 by an additional doublet at 4.46 ppm and two more singlets (2.045 and 2.06 ppm). The chemical shift data point to 70 N-acetylated metabolites in urine oligosaccharides with the structures β-GlcNAc-(1→2)-α-Man-(1-3)-[β-GlcNAc(1→2)- α-Man-(1-6)]-β-Man-(1→4)-GlcNAc for band 1 and β-GlcNAc-(1→2)-αMan-(1-3)-[β-GlcNAc-(1→4)][β-GlcNAc-(1→2)]-α-Man-(1-6)]-β-Man-(1→4)GlcNAc for band 2 [136]. The one-dimensional NMR spectrum for urine of an affected patient (Figure 5H) showed singlets between 2.02 and 2.08 ppm. The singlet with the highest intensity (2.06 ppm) represents a N-acetylated storage compound at a concentration of 77 µmol/mmol of creatinine. Sialidosis (OMIM 256550). Sialidosis, or mucolipidosis I, is caused by a deficiency of the lysosomal enzyme neuraminidase. The enzyme cleaves terminal α2→3 and α2→6 sialyl linkages of various glycopeptides, oligosaccharides, and gangliosides. High concentrations of sialyloligosaccharides (bound sialic acid) are found in urine of affected patients. At least six singlet resonances with a concentration >50 µmol/mmol of creatinine were found between 2.0 and 2.1 ppm by one-dimensional 1H-NMR spectroscopic analysis of the urine of our patient (Figure 5I). The COSY spectrum of the urine showed two cross-peaks: 1.72–2.67 ppm from the α2→6 sialyl linkages and 1.81– 2.67 ppm from the α2→3 sialyl linkages (data not shown). The singlet with the highest intensity (2.03 ppm) was present at a concentration of 430 µmol/mmol of creatinine. Using TLC, we preparatively isolated two oligosaccharides from the urine of an affected patient (Figure 1, lane 9). We observed two multiplets (1.75 and 2.66 ppm; cross-peak in the COSY spectrum) in the NMR spectra of both isolated compounds. This allowed us to conclude that both isolated compounds have an 2→6 sialyl linkage. In the spectrum for band 1, three N-acetyl singlets resonated at 2.02, 2.03, and 2.06 ppm, deriving from the two GlcNAc units and the acetylneuraminic acid (NeuAc) unit. From its spectrum we could conclude that band 1 probably is α-NeuAc-(2→6)-β-Gal-(1→4)-β-GlcNAc-(1→2)-α-Man(1→3)-β-Man-(1→4)-GlcNAc [137]. Band 2 differed in the chemical shift position of one acetyl singlet (2.05 instead of 2.03 ppm) and the presence of an additional signal at 4.11 ppm, which could indicate compound V in the report by Dorland et al. [137]. Canavan disease (OMIM 271900). Canavan disease is a leukodystrophy attributable to a deficiency of aspartoacylase (EC 3.5.1.15) and belongs to the group of neurotransmitter diseases. We analyzed urine from two unrelated patients. The one-dimensional 1H-NMR spectra for body fluid showed the presence of high concentrations of N-acetylaspartic acid [Figure 5J; urine from case A, 1760 µmol/mmol of creatinine; urine from case B, 1474 µmol/mmol of creatinine; reference values for all ages, 6–36 µmol/mmol of creatinine [138]]. 71 Chapter 2.1 Citrullinemia (OMIM 215700). Deficiency of the urea cycle enzyme argininosuccinate synthase (EC 6.3.4.5) leads to increased concentrations of citrulline and N-acetylcitrulline in the urine [124;139]. We analyzed two urine samples from a patient with citrullinemia. The urine 1HNMR spectrum showed citrulline multiplets (concentration, 825 µmol/mmol of creatinine). We assigned the unusually high singlet at 2.03 ppm to Nacetylcitrulline by comparing the chemical shifts with reference spectra in the literature [57]. N-Acetylcitrulline is not commercially available. The urinary Nacetylcitrulline concentrations were 441 and 518 µmol/mmol of creatinine (Figure 5K). Tyrosinemia type I (OMIM 276700). Tyrosinemia type I is caused by a deficiency of the enzyme fumarylacetoacetate hydrolase (EC 3.7.1.2). High body fluid concentrations of tyrosine, 4hydroxyphenylacetic acid, 4-hydroxyphenyllactic acid, succinylacetone, and methionine are characteristic of affected patients. We analyzed the urine of a 1year-old boy with this inborn error. In one-dimensional 1H-NMR spectra of his urine, we observed high concentrations of tyrosine and 4-hydroxyphenyllactic acid; the concentrations of betaine and methionine were also high. In addition to these typical tyrosinemia type 1 metabolites, we found a high concentration of N-acetyltyrosine (singlet at 1.94 ppm; 195 µmol/mmol of creatinine; Figure 5L). We identified the compound by comparing the spectrum with the model compound (singlet at 1.94 ppm; AB system at 3.07 ppm and AB system at 7.89 ppm). Succinylacetone and succinylacetoacetate, the biochemical hallmarks in tyrosinemia type I, could not be demonstrated by NMR spectroscopy in this urine. Tyrosinemia type II (OMIM 276600). Deficiency of the cytosolic enzyme tyrosine aminotransferase (EC 2.6.1.5) forms the primary defect of tyrosinemia type II. In the liver it leads to high concentrations of tyrosine, 4-hydroxyphenyllactic acid, 4-hydroxyphenylacetic acid, 4hydroxyphenylpyruvic acid, and N-acetyltyrosine in body fluids. Urine from one patient was analyzed. The concentration of N-acetyltyrosine in the urine of an affected patient was 95 µmol/mmol of creatinine (singlet at 1.94 ppm). In another patient (age, 0.2 years), we observed a much higher concentration of Nacetyltyrosine in the urine (>3000 µmol/mmol of creatinine). This patient received parenteral nutrition containing N-acetyltyrosine. In premature infants (gestational age at birth <36 weeks), urinary N-acetyltyrosine may be as high as 781 µmol/mmol of creatinine [140]. In addition, N-acetyltyrosine may be increased in some liver diseases, in which it may be observed along with N-acetyltryptophan (Table 1) 72 N-acetylated metabolites in urine Discussion N-Acetylated compounds are metabolites in many metabolic pathways and therefore may relate to various inborn errors of metabolism. In this study, we gave examples of 13 inherited diseases involving N-acetylated metabolites. At present, a comprehensive analytical technique to detect and identify N-acetylated compounds in body fluids is not available. However, N-acetylaspartic acid, the biochemical hallmark in Canavan disease, is detected by organic acid analysis by gas chromatography–mass spectrometry. Gas chromatography–mass spectrometry can also be used for the accurate diagnosis of tyrosinemia by analysis for all relevant metabolites, including N-acetyltyrosine. TLC is useful in the identification of abnormal N-acetylated oligosaccharides characteristic of some of the lysosomal storage disorders. NMR spectroscopy of body fluids can provide an overview of N-acetylated compounds. The protons of the N-acetyl group resonate in a specific narrow region of the spectrum. For metabolic screening purposes, we used the highest resonance in the N-acetyl region to identify samples with unusually high concentrations of an N-acetylated metabolite (Figure 3). This approach allowed different N-acetylated metabolites to be quantified in individual samples. For example, in the urine of a young child, the singlet at 1.94 ppm gave the highest concentration, whereas in an adult urine, the highest concentration was for the singlet at 2.04 ppm. As shown in Figure 3, most patients with inborn errors of metabolism excreted the N-acetylated metabolites typical for their diseases at concentrations >100 µmol/mmol of creatinine. In 10 of the 13 inborn errors of metabolism, this concentration was clearly higher than the upper limit of the reference interval. In all of these cases the combination of one-dimensional spectroscopy and COSY spectra allowed identification of the compound of interest. This led to the correct diagnosis in 9 of the 13 cases. Exceptions included the urines from patients with tyrosinemia type I and II and GM1- and GM2gangliosidosis. In tyrosinemia type I, the excretion of succinylacetone and succinylacetoacetate in urine is the diagnostic hallmark. To date, gas chromatography–mass spectrometry remains the method of choice to determine succinylacetone and the other characteristic metabolites and to diagnose tyrosinemia type I. For similar reasons the same holds true for tyrosinemia type II. Because of the relatively low urinary concentrations of the GM1- and GM2gangliosidosis metabolites (34 and 77 µmol/mmol of creatinine), assignment of these metabolites in urine in affected patients was not possible, and the diagnosis of GM1- and GM2-gangliosidosis was missed by 1H-NMR. NMR spectrometers with higher fields (>500 MHz) and spectrometers using the increased sensitivity of CryoProbe technology could probably detect these two inborn errors of metabolism. 73 Chapter 2.1 In spite of pH standardization, assignments of structure based solely on the chemical shift of the N-acetyl protons are insufficient because many acetyl singlets may occur in urine. Therefore, characteristic COSY patterns of model compounds are required to confirm the presence of individual N-acetylated metabolites in a patient’s urine. For example, the acetyl protons of N-acetylaspartic acid and Man(1-4)GlcNAc showed a singlet at the same chemical shift (2.03 ppm) in the one-dimensional spectrum. Two-dimensional COSY NMR experiments showed for each model compound different and characteristic cross-peaks produced by protons elsewhere in the molecule (Table 1). Because of a change in structure, bound N-acetylneuraminic acid can be distinguished from free N-acetylneuraminic acid in two-dimensional COSY 1H-NMR spectra (Table 1). This makes it possible to differentiate patients with isolated neuraminidase deficiency, who excrete bound N-acetylneuraminic acid, from patients with Salla disease or French type sialuria, who excrete excess free N-acetylneuraminic acid in urine [34]. However, in regions with a high density of resonances, especially around the two-dimensional COSY diagonal, cross-peaks sometimes show overlap and cannot be used for correct assignments. Other two-dimensional NMR techniques, such as J-resolved spectroscopy and total correlation spectroscopy, can be used to help with further interpretation. The use of medications can be a pitfall in the quantification of N-acetylated metabolites. The antiepileptic drug Sabril® (vigabatrin) is often used in patients suspected to have an inborn error of metabolism. The 1H-NMR urine spectra of Sabril-treated patients show high Sabril resonances between 2.0 and 2.1 ppm caused by CH2 protons (multiplet). This Sabril multiplet interferes with the N-acetyl resonances. Therefore, correct interpretation and quantification of the N-acetyl region of the 1H-NMR spectrum is impossible. Some drugs can be converted in the body to their acetylated conjugates. Acetaminophen, for example, is metabolized in the body and is excreted in the urine in various forms. The 1H-NMR spectrum of urine may show acetaminophen (paracetamol; singlet at 2.14 ppm), 4glucuronosidoacetanilide (singlet at 2.15 ppm), N-acetyl-4-aminophenol sulfate (singlet at 2.17 ppm), and N-acetyl-2-(N-acetyl-L-cysteinyl)-4-aminophenol [120]. In conclusion, we have shown that careful inspection and quantitative interpretation of the N-acetyl region of the 1H-NMR spectrum may identify an abnormally high concentration of an N-acetylated metabolite. The technique is uniquely suited for studying patients with inborn errors of metabolism, and our data extend that to diseases involving N-acetylated metabolites. A combination of one- and two-dimensional COSY 1H-NMR spectroscopy identified at least nine different inborn errors of metabolism involving N-acetylated metabolites. 74 N-acetylated metabolites in urine Acknowledgements The 1H-NMR spectra were recorded at the Dutch hf-NMR facility at the Department of Biophysical Chemistry, University of Nijmegen, The Netherlands (Prof. Sybren Wijmenga, department head). We thank Jos Joordens for invaluable help and technical assistance. We also thank Drs. Neidig and Fischer (Bruker BioSpin, Karlsruhe, Germany) for invaluable help and improvements in the AMIX software. 75 76 CHAPTER 2.2 Severe hypomyelination associated with increased levels of N-acetylaspartylglutamate in CSF Nicole I. Wolf, Michèl A.A.P. Willemsen, Udo F.H. Engelke, Marjo S. van der Knaap, Petra J.W. Pouwels, Inga Harting, Johannes Zschocke, Erik A. Sistermans, Dietz Rating and Ron A. Wevers NEUROLOGY 2004; 62: 1503-1508 77 78 N-acetylaspartylglutamate in CSF Abstract Background: Two unrelated girls had early onset of nystagmus and epilepsy, absent psychomotor development, and almost complete cerebral MRI. The clinical features and MR images of both connatal form of Pelizaeus–Merzbacher disease (PMD), recessive disorder caused by duplications or mutations of gene (PLP). absence of myelin on patients resembled the which is an X-linked the proteolipid protein Objective: To define a unique neurometabolic disorder with failure of myelination. Methods and Results: 1H-NMR of CSF in both girls was performed repeatedly, and both showed highly elevated concentrations of N-acetylaspartylglutamate (NAAG). The coding sequence of the gene coding for glutamate carboxypeptidase II, which converts NAAG to N-acetylaspartate (NAA) and glutamate, was entirely sequenced but revealed no mutations. Even though both patients are girls, the authors sequenced the PLP gene and found no abnormality. Conclusions: NAAG is an abundant peptide neurotransmitter whose exact role is unclear. NAAG is implicated in two cases of unresolved severe CNS disorder. Its elevated concentration in CSF may be the biochemical hallmark for a novel neurometabolic disorder. The cause of its accumulation is still unclear. 79 Chapter 2.2 Introduction White matter disorders of infancy and childhood may cause severe neurologic disability. The disorders can be divided into demyelinating and hypomyelinating processes, depending on whether normal myelin is formed and subsequently degraded or whether the process of normal myelin formation never or incompletely occurs [141;142]. The differential diagnosis of white matter disorders represents a major challenge, leaving many patients without a specific diagnosis. There have been some efforts to further classify unsolved white matter disorders. One approach takes typical changes of MRI into account to categorize MRI findings systematically and thus, by recognizing typical patterns, to delineate new disease entities [143]. New techniques such as diffusion-weighted imaging [144] and in vivo 1H-NMR spectroscopy further differentiate white matter disorders. The latter has been used in the elucidation of several new neurometabolic disorders [70;145;146]. Another powerful tool in the investigation of white matter disorders is in vitro 1H-NMR spectroscopy of CSF. It is able to detect a wide range of metabolites and also has contributed to the delineation of new inherited inborn errors of metabolism [33;36;42]. In this report, we describe two unrelated girls presenting with nystagmus, intractable epilepsy, and severe psychomotor retardation. MRI showed almost absent myelination of the brain. In vitro 1H-NMR spectroscopy of CSF revealed strongly elevated N-acetylaspartylglutamate (NAAG). The phenotype resembles Pelizaeus–Merzbacher disease (PMD) [147], which occurs in boys and is caused by alterations of the proteolipid protein gene (PLP) on the X-chromosome [148]. However, PLP gene alterations were ruled out. We speculate that the marked elevation of NAAG in CSF may indicate an unrecognized neurodegenerative disorder, but the underlying genetic defect is still unclear. Patients and Methods CASE REPORTS Patient 1 is an 8-year old girl, born after an uneventful pregnancy as the second child of unrelated, healthy parents. The first symptom at age 6 weeks was a fine rotatory nystagmus. At age 5 months, severe hypotonia was noted. Head growth decelerated during the first 6 months of life. Epilepsy with secondarily generalized tonic seizures was overt from age 24 months and resistant to therapy. The initial hypotonia evolved into a severe spastic paraparesis with positive Babinski signs, which was most prominent at age 2 years. Now, at age 8 years, the patient is again hypotonic with some spastic features (scissoring of the legs). Tendon reflexes were initially normal, at the age of 12 months clearly exaggerated, but disappeared around the age of 5 years, suggesting the development of a 80 N-acetylaspartylglutamate in CSF peripheral neuropathy. Feeding difficulties are prominent. An older sister has slight mental retardation without abnormalities on cerebral MRI. Two younger sisters are healthy. Parents declined further investigations in this child. Patient 2 is a 3-year-old girl from healthy and unrelated parents. She has a healthy older brother. At age 4 weeks, the parents noted head nodding and nystagmus. At age 6 weeks, seizures started, which could not be controlled. Microcephaly developed during the first 4 months. A gastric tube was inserted at the age of 2 months because of feeding difficulties, but the child failed to thrive. Neurologic examination revealed severe hypotonia, no head control, and absent tendon reflexes. No developmental milestones were achieved. The course of the disease was further complicated by severe nonepileptic myoclonus and dystonia. MRI The first cerebral MRI of Patient 1 was performed at the age of 2 years, the second at age 3.5 years (not shown), and the third at age 5 years. The cerebral MRI of Patient 2 were obtained at ages 11 weeks and 2, 3, and 3.5 years. The imaging protocol for the brain comprised sagittal and transverse T1-weighted (repetition time [TR] 570 milliseconds, echo time [TE] 14 milliseconds) and T2-weighted (TR 3,000 milliseconds, TE 22, 60, and 120 milliseconds) spin echo imaging and coronal or transverse FLAIR (fluid-attenuated inversion recovery) images (TR 9,000 milliseconds, TE 105 milliseconds, inversion time 2,200 milliseconds). IN VITRO 1H-NMR SPECTROSCOPY OF CSF AND URINE Sample preparation. After deproteinization using a 10-kd filter (Sartorius, Goettingen, Germany), 20.2 mM trimethylsilyl-2,2,2,3- tetradeuteropropionic acid (TSP; sodium salt; Merck, Whitehouse Station, NJ) was added to the samples for spectrum calibration and metabolite quantification. Final TSP concentration was 0.8 mM in CSF and 1.8 mM in urine. The pH was adjusted to 2.50 ± 0.05. NAAG as model compound was commercially available from Sigma (A5930; St. Louis, MO). One-dimensional 1H-NMR spectroscopy. Single-pulse experiments were recorded using a 60° pulse at 298 K on a Bruker AMX-500 spectrometer (Bruker Analytische Messtechnik, Zurich, Switzerland), a 5mm TXI 1H/15N/13C xyz gradient probe head, and 5-mm Wilmad Royal Imperial NMR tubes (Aldrich, Milwaukee, WI). The temperature was set at 298 K. The sweep width was 6002 Hz. For urine, 128 scans were recorded and for CSF 256 scans. Per scan, 32K data points were sampled. The relaxation delay was 10 seconds, during which the water resonance was presaturated. The spectra were analyzed using one-dimensional WinNMR and AMIX software (Bruker Analytische Messtechnik). A sine-bell-squared filter (SSB = 2) was used, the spectra were zero-filled to 64K data points, and Fourier transformation was performed. Phase and baseline were 81 Chapter 2.2 corrected manually. The concentrations were calculated by comparing the peak areas of the metabolites with those of creatinine in urine or TSP in CSF. 1H-1H correlation spectroscopy. The spectra were recorded at 500 MHz using 4K data points in F2 and a spectral width of 6,002 Hz. For the CSF samples, 256 increments and 16 scans/increment were used. The TR was 1.5 seconds, during which the water resonance was presaturated. Prior to Fourier transformation, a sine function was applied in both time domains. IN VIVO 1H-NMR SPECTROSCOPY Patient 2 underwent singlevoxel MR spectroscopy at age 3.5 years, as described previously [70]. Volumes of interest (VOI) were selected within the “white matter” of the centrum semiovale in both hemispheres (4 mL) and the midparietal cortex (8 mL). In each VOI, a fully relaxed, short echo time (STEAM) spectrum (TR/TE/mixing time = 6,000/20/10 milliseconds, 64 accumulations) was obtained. Metabolite concentrations were calculated using LCModel [149;150] and expressed as millimoles per liter of VOI. Data were compared with values obtained in agematched control subjects [151]. Figure 1 NAA and NAAG signaling and metabolism in the CNS: 1 = NAA synthase 2 = NAAG synthase 4 = glutamate carboxypeptidase II 3 NAA NAAG = N-acetylaspartate = N-acetylaspartylglutamate AcCoA = acetyl-coenzyme A Ac = acetate Glu Asp 82 = aspartoacylase = glutamate = aspartate N-acetylaspartylglutamate in CSF MUTATION ANALYSIS OF PLP AND FOLH1 GENES Patient genomic DNA was extracted from leukocytes using standard procedures. PLP gene analysis (amplification of individual exons and quantitative Southern blotting) was done as previously described, with the exception that exon 1 was not analyzed by restriction fragment length polymorphism analysis but by sequencing analysis [152]. The FOLH1 gene (GenBank accession no. AF007544) encodes for the enzyme glutamate carboxypeptidase II (GCP II; EC 3.4.17.21) involved in NAAG catabolism (Figure 1). For analyzing this gene, the 19 exons and exon–intron boundaries were amplified separately. Primer sequences used are available on request. Each primer contained an additional 17mer corresponding to a universal recognition sequence (for the forward primers P172: TATAGGGCGAATTGGGT; for the reverse primers M13: GTAAAACGACGGCCAGT). After purification, the PCR products were cycle sequenced with the appropriate 5'-labeled universal primers using a dye terminator sequencing kit (Applied Biosystems, Foster City, CA). Samples were analyzed on an ABI 3100 automatic sequencer (Applied Biosystems). Results LABORATORY AND ELECTROPHYSIOLOGIC STUDIES. Routine blood hematology and chemistry panels were normal. Urinary excretion of organic acids and amino acids, oligosaccharides, and free and bound neuraminic acid was normal. Plasma amino acids, very long chain fatty acids, and phytanic acid were normal. Isoelectric focusing of serum transferring gave a normal pattern. Activities of lysosomal enzymes measured in leukocytes were normal. CSF glucose, cells, albumin, and protein levels were within normal limits. Lactate in blood and CSF was not increased. Still, a quadriceps muscle biopsy was performed in Patient 2 and gave normal results for muscle histology. Activities of the individual complexes of the respiratory chain were normal. Chromosomal analysis showed a normal 46, XX karyotype in both girls. Electrophysiologic studies were carried out at age 5 years, after the disappearance of tendon reflexes, in Patient 1 and at age 5 months in Patient 2. Motor nerve conduction velocity was normal in both, but sensory nerve conduction studies could not be interpreted properly in Patient 1. Brainstem evoked auditory potentials were abnormal in both children, showing normal wave I and absent waves II to V. MRI. With MRI of the brain performed on several occasions, hardly any myelin could be detected, as indicated by the homogeneous high signal of white matter on T2weighted images and low signal on T1-weighted images. The only evidence of myelin deposition was in the brainstem (Figure 2). The corpus callosum was very 83 Chapter 2.2 thin in both patients. The signals of cortical structures and deep gray nuclei were normal. During the follow-up period, Patient 1 displayed atrophy of supratentorial structures and to a lesser extent of the cerebellum. Patient 2 showed only mild supratentorial atrophy and a normal cerebellum without change over time. BODY FLUID 1H-NMR SPECTROSCOPY. Proton 1H-NMR spectroscopy of CSF of both patients (Figure 3) revealed an unknown singlet resonance at 2.04 ppm. Smaller resonances at 2.85 (AB system), 2.01, 2.23, 2.48, 4.44, and 4.72 ppm (multiplet resonances) deriving from the same molecule helped to assign the responsible metabolite as NAAG in onedimensional and two-dimensional correlation spectroscopy (COSY) spectra. Its identity was confirmed by measuring NAAG as model compound. Figure 3 shows the prominent 2.04 singlet resonance deriving from the N-acetyl moiety of the molecule and minor resonances around 2.85 ppm (COSY spectrum not shown). The concentration of NAAG repeatedly was strongly elevated in both patients (Patient 1: 95 and 90 µmol/l; Patient 2: 197 and 146 µmol/l; reference 4.4 ± 1.7 µmol/l [1 SD] [153]). CSF N-acetylaspartic acid (NAA) in our patients was consistently lower than 5 µmol/l (the in vitro NMR spectroscopy detection limit), which is considered normal. An elevated concentration of NAAG could not be demonstrated in the 1H-NMR spectrum of urine of our patients (not shown; detection limit 50 µmol/mol of creatinine). Using 1H-NMR spectroscopy, we never found NAAG levels higher than 12 µmol/l in a control group of 200 patients clinically suspected to have a neurometabolic disorder. Additionally, we examined CSF from patients both with defined (n = 16) and undefined (n = 5) disorders of the white matter. The group of patients with a defined white matter disease consisted of three children with X-linked adrenoleukodystrophy, two children with Krabbe disease, one child with metachromatic leukodystrophy, three children with Sjögren–Larsson syndrome (SLS), two children with vanishing white matter disease, one child with Aicardi–Goutières syndrome, two children with a mitochondrial disorder, and two children with Canavan disease. The highest NAAG level found was 12 µmol/l in one of the patients with SLS. Among the patients with unclassified white matter disease were three children with severe myelination delay (all of them with severe epilepsy), one child with hypomyelination, cerebellar atrophy, and progressive ataxia, and one child with elevated signal of the central white matter, progressive spastic paraplegia, and mental retardation. The NAAG concentrations in this group ranged from 2 to 11 µmol/l. IN VIVO 1H-NMR SPECTROSCOPY. In vivo 1H-NMR spectroscopy of white matter in Patient 2 showed an elevation of absolute concentrations of the sum of NAA and NAAG (tNAA) and also of creatine and phosphocreatine (tCr), resulting in a normal tNAA/tCr ratio. Choline was decreased in white matter. In parietal gray matter, the absolute tNAA 84 N-acetylaspartylglutamate in CSF concentration was elevated as well, but less than in white matter; the concentration ratio with tCr was normal (tables 1 and 2). Myo-inositol was normal in gray matter but increased in white matter. Table 1: Absolute metabolite concentrations Metabolites Cortex Patient White Matter Controls Patient Controls tCr 7.1 6.4 ± 0.7 8.2 ± 0.1 5.0 ± 0.4 tNAA 9.4 8.2 ± 0.9 11.1 ± 0.5 7.2 ± 0.7 Cho 1.2 1.1 ± 0.2 1.3 ± 0.1 1.9 ± 0.2 Ins 4.3 4.7 ± 0.4 4.8 ± 0.1 3.7 ± 0.6 In vivo 1H-MR spectroscopy findings in Patient 2 Values are absolute concentrations in mmol/l. * Control values are absolute concentrations expressed as means ± SD. Patient values in white matter are also means ± SD obtained from two white matter spectra of right and left parietal white matter. Cortex spectra were recorded in the parietal cortex and white matter spectra in right and left parietal white matter. tCR = creatine + phosphocreatine; tNAA = N-acetylaspartate + N-acetylaspartylglutamate; Cho = choline-containing compounds; Ins = myo-inositol. Table 2: Concentration ratios relative to tCr Ratios Cortex White Matter Patient Controls Patient Controls tNAA / tCr 1.33 1.28 ± 0.11 1.35 ± 0.04 1.45 ± 0.15 Cho / tCr 0.17 0.18 ± 0.02 0.16 ± 0.01 0.38 ± 0.03 Ins / tCr 0.61 0.73 ± 0.07 0.58 ± 0.00 0.74 ± 0.09 In vivo 1H-MR spectroscopy findings in Patient 2 Values are concentration ratios related to tCr. * Control values are expressed as means ± SD. Patient values in white matter are also means ± SD obtained from two white matter spectra of right and left parietal white matter. Cortex spectra were recorded in the parietal cortex and white matter spectra in right and left parietal white matter. tCR = creatine + phosphocreatine; tNAA = N-acetylaspartate + N-acetylaspartylglutamate; Cho = choline-containing compounds; Ins = myo-inositol. PLP/FOLH1 Analysis of the PLP gene did not reveal either duplication of the whole gene or other mutations in the coding region and the exon–intron boundaries. Sequencing of the FOLH1 gene revealed a heterozygous polymorphism (c222C3T resulting in H75Y) in exon 2 of Patient 1, but no pathogenic mutation. This polymorphism was observed in 23 of 98 alleles from healthy volunteers (data not shown). 85 Chapter 2.2 Discussion In this study, we describe two unrelated patients with almost absent CNS myelin, intractable epilepsy, nystagmus, progressive microcephaly, and severe psychomotor retardation. The phenotype could be compatible with PMD. Also, the MRI appearance could be consistent with this diagnosis, as the typical appearance of imaging studies in this disorder includes diffuse severe hypomyelination with a homogeneous high signal in T2-weighted images [154-156]. However, PMD is an X-linked recessive disorder, and PLP gene alterations were ruled out in both patients. We consider the markedly elevated concentrations of NAAG in CSF, found by in vitro 1H-NMR spectroscopy, as the biochemical hallmark of a novel neurodegenerative disorder in these patients. Several girls with symptoms similar to the severe connatal form of PMD but without PLP alterations have been described. This was interpreted as an autosomal recessive form of the disease with a still unresolved genetic basis [157;158]. Therefore, the connatal PMD phenotype without PLP alterations may represent a separate genetic entity. Sequencing of the whole PLP gene (exons and intron–exon boundaries) was performed in both patients and revealed neither point mutations nor duplication. Female carriers of PLP gene alterations are usually asymptomatic. A few mildly affected female patients with PLP gene alterations (including duplications and null mutations) have been described [159-162]. In some families, they develop late-onset spastic paraplegia [163;164]. The clinical and molecular genetic data in our patients thus make PMD due to PLP gene alterations highly unlikely. In vivo 1H-NMR spectroscopy of one of our patients revealed elevation of absolute levels of tNAA in white and gray matter. In relation to creatine, tNAA levels were normal. Choline was significantly reduced in white matter. Similar findings [165] as well as conflicting results [155;166-169] have been reported in PMD patients. The elevation of tNAA in white matter in our patients may be caused by a denser packing of axons secondary to the lack of myelin sheets, thus indicating the presence of intact axons in the absence of myelin. Alternatively, a high concentration of NAAG could explain the tNAA elevation. The N-acetyl group of NAAG resonates at 2.04 ppm and of NAA at 2.02 ppm. Because of its high resolution, in vitro 1H-NMR spectroscopy is able to distinguish these resonances, but this may often be difficult in in vivo 1H-NMR spectroscopy. In our white matter spectra, however, a prominent NAAG signal could not be detected in spite of good spectral resolution. 86 N-acetylaspartylglutamate in CSF Figure 2 Cerebral MRI of Patient 1 (A) and Patient 2 (B), demonstrating almost complete absence of myelin in supra- and infratentorial structures. Gray matter (cortex and deep nuclei) appears normal. The corpus callosum is string-like. Note the supratentorial and cerebellar atrophy in Patient 1. Axial pictures are T2 weighted, sagittal pictures T1 weighted. 87 Chapter 2.2 As our data on other leukodystrophies show, the clear elevation of NAAG in our two patients is not a nonspecific finding associated with white matter disorders in general. A mild increase in the concentration of NAAG in CSF has been described in combination with highly elevated concentrations of NAA in a case of Canavan disease [170]. Figure 3 spectroscopy of CSF: (A) CSF of patient 1 with a strong N-acetyl resonance at 2.04 ppm and smaller resonances around 2.85 ppm. (B) normal control. 1H-NMR During a recent congress, NAAG levels were also reported to be elevated in body fluids from boys with PMD [171]. However, the abstract gives details neither on the clinical and genetic characteristics nor on the concentration of NAAG in body fluids of the patients. There is no explanation as yet for this finding. Further data on the NAAG concentration in body fluids of patients with PLP alterations are pivotal in unraveling the pathophysiologic mechanisms underlying this phenomenon. NAAG is one of the most abundant metabolites in vertebrate brain [172;173]. It is synthesized by the enzyme NAAG synthase (N-acetylaspartate-L-glutamate ligase) from NAA and glutamate and degraded by GCP II, formerly called NAALADase (Figure 1) [174;175]. In the brain, the latter enzyme is exclusively expressed on the membrane surface of astrocytes. NAA is thought to be involved in neuron-toglia signaling; NAAG is targeted to the astrocyte and thus might contribute to neuron-to-astrocyte signaling, which might be important for processes as myelination and maintenance of axons [173]. Based on the above, we 88 N-acetylaspartylglutamate in CSF hypothesized that accumulation of NAAG in our patients could be caused by deficiency of the NAAG-degrading enzyme, GCP II. Sequencing the coding region of the corresponding FOLH1 gene, however, did not reveal pathogenic mutations. Thus, GCP II deficiency is not likely for the NAAG accumulation in our patients, although promoter mutations of the gene have not been ruled out. Very recently, GCP II knock-out mice were found to exhibit no abnormalities of the CNS, and NAAG degradation was normal in the brain tissue. The finding was explained by the presence of another enzyme with NAAG peptidase capacity [176]. In humans as well, there are enzymes with NAAG hydrolyzing activity besides GCP II, but it is not known whether they are expressed in the CNS [177]. Impaired function of one of those enzymes or a defect in the complex NAAG transport and receptor system (Figure 1) could theoretically cause NAAG elevation in the CNS of our patients. Another explanation could be that abnormal, severely reduced, or absent myelin impairs the normal function of the NAAG degrading enzyme on the astrocyte membranes and leads to secondary elevation of this metabolite. Elevated NAAG in boys with PLP alterations might favor this hypothesis. NAAG metabolism and molecular genetic studies of the genes involved should be a focus for further pathophysiologic studies in CNS disorders with severe hypomyelination. 89 90 CHAPTER 2.3 NMR spectroscopy of aminoacylase 1 deficiency, a novel inborn error of metabolism Udo F.H. Engelke, Jörn Oliver Sass, Rudy N. Van Coster, Erik Gerlo, Heike Olbrich, Stefan Krywawych, Jacqui Calvin, Claire Hart, Heymut Omran and Ron A. Wevers NMR in Biomedicine 2007; in press 91 92 Aminoacylase 1 deficiency Abstract Aminoacylase 1 deficiency is a novel inborn error of metabolism. The clinical significance of the deficiency is under discussion, as well as the possible consequences of the defect for brain metabolism and function. This study includes the five originally published cases as well as three novel ones. NMR spectroscopy of urine, serum and cerebrospinal fluid has been used to study these patients. A typical profile with 11 accumulating N-acetylated amino acids was observed in urine from the patients. The concentration of most of the accumulating metabolites is typically 100–500 µmol/mmol creatinine. Two additional minor N- acetylated metabolites accumulating remain metabolites are unidentified. <20 µmol/L The in concentrations serum from the of the patients. Interestingly we found no evidence of an increased concentration of N-acetylated amino acids in the cerebrospinal fluid from one patient. Our data define aminoacylase 1 deficiency at the metabolite level providing a specific urinary profile of accumulating N-acetylated amino acids. 93 Chapter 2.3 INTRODUCTION Proton NMR spectroscopy has been successfully applied to the field of inborn errors of metabolism (IEM) [9]. The technique can be used for rapid identification and quantification of metabolites in body fluids. Known, but also as yet unknown, IEM have been diagnosed using NMR spectroscopy of body fluids [33;36;42]. Aminoacylase 1 deficiency is a novel inborn error of metabolism (OMIM: 609924). In two recent papers, clinical and laboratory data from 5 children with excessive urinary excretion of various N-acetylated amino acids have been described [54;55]. The patients were found to have a deficiency of aminoacylase 1 (ACY1; EC 3.5.1.14) which catalyzes the deacetylation of N-acetylamino acids. In this paper, we provide a more detailed description of the NMR spectroscopic findings of the five original patients with ACY1 deficiency as well as three novel cases. Four types of aminoacylase have been identified on the basis of substrate specificity [178]. ACY1 is encoded by the gene aminoacylase 1 (ACY1) [54]. The enzyme preferentially uses neutral aliphatic amino acids with a short-chain acyl moiety as substrate [178]. These include the N-acetylated forms of serine, threonine, alanine, methionine and glycine. The enzyme aminoacylase 2 (ACY2) or aspartoacylase (EC 3.5.1.15) catalyzes the deacetylation of N-acetylaspartic acid (NAA) into aspartic acid. Deficiency of this enzyme causes Canavan disease (OMIM 271900) and results in high concentrations of NAA in body fluids of affected patients. Aminoacylase 3 (ACY3) acts preferentially on N-acylaromatic amino acids and the enzyme acyl-lysine deacylase catalyzes the N-deacetylation of N5acetyllysine. A deficiency of ACY3 and acyl-lysine deacylase in humans has not been reported. In this paper, we present data at the metabolite level typical for ACY1 deficiency and show that NMR spectroscopy can easily discriminate between ACY1 and ACY2 deficiency. Moreover, this study presents the NMR characteristics of a broad range of N-acetylated amino acids relevant for the diagnosis of several other inborn errors of metabolism [179]. MATERIALS AND METHODS PATIENTS. Eight unrelated patients with ACY1 deficiency were included in this study (Table 1). Clinical details about patient 1 [55] and patients 2 – 5 (cases OS-104 II-1, OS-127 II-1, OS-124 II-1, OS-120 II-1 in [54], respectively) have been described elsewhere. Patient 6 and 8 are children from independent families of English and Pakistani ancestry. The case histories will be described separately (cases OS138II-1 94 Aminoacylase 1 deficiency and OS-152II-1 [180]). Patient 7 is the third child of distantly related Pakistani parents. His father, older brother and sister are well. His mother has learning difficulties, apparently acquired at the age of 8 years, following an illness. At the age of 13 weeks there were concerns that he was not fixing and following, thought to be consistent with delayed visual maturation. At the age of 5 months he was noted to be very floppy and was referred for investigation of global developmental delay and failure to thrive. At this time he was found to have abnormal acetylated amino acids in the urine with normal plasma amino acids and blood spot acylcarnitine profile. Currently, at the age of 21 months, he has severe developmental delay and failure to thrive (weight << 2nd centile). He remains profoundly hypotonic with extremely poor head control. The delayed visual maturation is improving although he now suffers from intermittent nystagmus. The Canavan patient was a 1-year-old boy with neurological involvement, mental retardation, demyelination and loss of acquired skills. THE DIAGNOSIS. In most patients, the diagnosis ACY1 deficiency (OMIM 609924) was confirmed at the metabolite level (GC-MS and NMR spectroscopy of urine), at the enzyme level (ACY1 deficiency in EBV transformed lymphocytes) and the molecular genetic level. In patient 5 the defect was established at the metabolite and the molecular genetic level only. In patient 7 the defect was established at the metabolite and the enzyme level. The diagnosis in the Canavan patient (ACY2 deficiency) was established on basis of the clinical features and observations at the metabolite level (GC-MS and NMR spectroscopy of body fluids). Further confirmation was obtained by enzymatic analysis in fibroblasts. AUTHENTIC STANDARDS. Most N-acetylated amino acids, listed in Table 2 and 3, were commercially available and were purchased at Sigma or Aldrich. Exceptions included Nacetylthreonine, N-acetylcitrulline, N-acetyltaurine and N-acetylisoleucine. These compounds were synthesized by Chiralix (Chiralix B.V., Toernooiveld 100, 6525 EC Nijmegen, The Netherlands), by acylation of the corresponding amino acid with acetic acid anhydride in 1M aqueous sodium hydroxide [181]. SAMPLE PREPARATION. All body fluids were centrifuged for cell separation and stored at low temperature (-80°C for CSF; -20°C for urine and plasma) until NMR analysis. Plasma and CSF samples were deproteinised using Sartorius Centrisart I filters (cut-off 10kDa; Sartorius AG, Göttingen, Germany). Before use, the filter was washed twice by centrifugation of water to remove glycerol. Urine was centrifuged before analysis. An aliquot (20 µl for plasma and CSF; 70 µl for urine) of 20.2 mM trimethylsilyl2,2,3,3-tetradeuteropropionic acid (TSP, sodium salt; Aldrich) in D2O was added 95 120 90 70 150 310 160 160 360 250 290 120 150 160 nd nd 2 (13.2) 2 (13.4) 3 (1.4) 3 (1.6) 4 (0.6) 5 (3.1) 6 (3.3) 7 (0.9) 7 (0.7) 7 (0.5) 8 (0.93) 8 (0.95) 8 (1.43) Canavan Ref[182] nd nd + + + + + + + + + + + + + + + Leucine N-Ac. nd nd ++ ++ ++ + + + + + ++ + + + + + ++ Asparagine N-Ac. nd nd 270 290 290 540 490 590 250 280 740 230 280 170 190 190 510 Glutamic acid N-Ac. nd nd ++ ++ ++ + + ++ + + ++ + + + + + ++ Glutamine N-Ac. nd nd 120 50 50 nd nd nd + nd nd 580 590 300 260 + Nd + 350 + b Nd 420 80 520 430 300 180 90 130 110 + + + + + + + + + 230 Glycine Isoleucine + N-Ac. N-Ac. b a 160 60 50 80 80 80 50 50 40 120 Methionine N-Ac. nd nd + + + + + + + + + + + + + + + Valine N-Ac. Artifact Not detectable (< ± 5 µmol/mmol creatinine b nd + / ++ Semi-quantitative estimate from the 1D 1H NMR spectrum (+=10 – 250, and ++ = 250 - 500 µmol/mmol creatinine) Quantification on S-CH3 singlet (Figure 1; peak 12) 90 2 (13.1) a 200 Alanine (age) 1 (0.3) N-Ac. Patient nr. Canavan disease nd nd 220 210 220 280 260 350 160 210 390 150 <10 100 110 100 330 b Serine N-Ac. nd Nd 30 20 30 60 40 50 20 30 40 <40 1470 <10 <10 <10 <10 <10 <10 <10 <10 <10 <10 <10 <10 20 <10 b <10 <10 <10 NAA 20 20 20 50 Threonine N-Ac. Table 1. Urine concentrations (μmol/mmol creatinine) of the N-acetylated amino acids from patients with ACY1 deficiency (1 – 8) and the patient with Aminoacylase 1 deficiency to 700 µL of the ultrafiltrate or urine, providing a chemical shift reference (δH = 0.00 ppm), a concentration reference and a deuterium lock signal. The pH of each sample was adjusted to 2.50 ± 0.05 with concentrated HCl. Finally, 650 µL of the sample was placed in a 5mm NMR tube (Wilmad Royal Imperial). ONE-DIMENSIONAL 1H-NMR SPECTROSCOPY. Body fluids samples were measured at 500 MHz on a Bruker DRX 500 spectrometer essentially as described before [183]. The acetyl methyl protons of N-acetylated compounds have a T1 relaxation time of approximately 2 s [11]. Therefore, we used a recycle time of 10 s, which is 5 times the T1 value of Nacetylated metabolites. The N-acetyl resonances in the 1-dimensional urine spectra were fitted semiautomatically to a Lorentzian line shape model function. The integrals of these fits were used for quantification by comparing them to the fit integral of creatinine (singlet: 3.13 ppm) and concentrations were expressed as µmol/mmol creatinine. If computer fitting was not possible on these singlets due to severe overlap we estimated concentration ranges based on peak intensity of the N-acetyl CH3 resonances and the creatinine CH3 resonance (Table 1: + / ++ semiquantitative estimate from the 1D 1H-NMR spectra). For CSF, the N-acetylated amino acid resonances and the TSP singlet (nine equivalent protons) were fitted semi-automatically with Lorentzian line shapes and concentrations were calculated from the relative integrals of the fitted lineshapes using the known concentration of TSP. Concentrations were expressed as µmol/liter. For identification and quantification of the NMR spectra, Topspin and AMIX software were used (Bruker BioSpin, Karlsruhe, Germany). 1H-1H COSY SPECTROSCOPY. 2-dimensional COSY NMR urine spectra and all model compound spectra were measured at 500 MHz on a Bruker DRX 500 spectrometer essentially as described before [47]. 1H-13C HSQC SPECTROSCOPY. Three HSQC urine spectra from different ACY1 patients (Patients 1, 2 and 6) were measured at 500 MHz on a Bruker AV 500 spectrometer using a triple-resonance inverse CryoProbeTM (Bruker), wherein the preamplifier and the RF coils are cooled to low temperature. Furthermore, the HSQC spectra were measured of the model coumpounds N-acetylalanine, Nacetylleucine, N-acetylasparagine, N-acetylglutamic acid, N-acetylglutamine, Nacetylmethionine, N-acetylisoleucine, N-acetylglycine, N-acetylvaline, N-acetylserine and N-acetylthreonine. Details about the measurement have been described elsewhere [47]. GAS CHROMATOGRAPHY MASS-SPECTROSCOPY. N-Acetyl amino acids were determined with GC-MS after liquid extraction from urine and derivatization with BSTFA [55;182]) and/or diazomethane [54;184]). 97 Chapter 2.3 RESULTS We analyzed 15 urine samples from eight patients with ACY1 deficiency (Table 1). In Figure 1A, the urine 1H-NMR spectrum (1.3 – 3.3 ppm) from patient 6 is shown. The major metabolites in this region and the chemical shifts at pH 2.5 are creatinine (singlet: 3.13 ppm), creatine (singlet: 3.05 ppm), citric acid ([AB]2 spin system: 2.99; 2.83 ppm) and alanine (doublet: 1.51 ppm). Furthermore, several unusually high singlet resonances were observed in the region between 2.0 and 2.2 ppm (Figure 1A; insert). In this region, CH3 protons from N-acetyl groups characteristically give a singlet resonance. In control urine samples, the highest concentration of an N-acetylated metabolite is generally below 40 µmol/mmol creatinine [11]. In the urine of patient 6 (Figure 1A), seven singlet resonances with a concentration higher than 40 µmol/mmol creatinine were observed in this Nacetyl region. This pattern was found in the urine from all 8 patients with ACY1 deficiency. NMR measurements of model compounds confirmed that these singlet resonances were caused by different N-acetylated amino acids. Resonances were assigned from N-acetylalanine, N-acetylleucine, N-acetylasparagine, Nacetylglutamic acid, N-acetylglutamine, N-acetylmethionine, N-acetylisoleucine, N-acetylglycine, N-acetylvaline, N-acetylserine and N-acetylthreonine (Table 2). Table 1 gives the quantitative values for the N-acetylated amino acids. IDENTIFICATION OF N-ACETYLATED AMINO ACIDS In addition to the unusual singlets in the N-acetyl region, two triplets at 2.49 and 2.38 ppm and an unknown doublet at 1.41 ppm were observed in the onedimensional 1H NMR urine spectrum (Figure 1A; peak 14, 13 and 15, respectively). These resonances were observed in the urine from all patients with ACY1 deficiency and are normally not present in urine spectra from healthy individuals. 1H-NMR of model compounds revealed that the doublet and singlet at 1.41 and 2.02 ppm, observed with an integral ratio 1:1 for the patient’s urine derive from N-acetylalanine. This was confirmed in a spiking experiment. Additional confirmation was provided by 2-dimensional NMR experiments (1H-1H COSY and 1H-13C HSQC) performed on an authentic reference solution of N-acetylalanine and the patient’s urine. The characteristic COSY and HSQC patterns of N-acetylalanine were observed in the 2-dimensional NMR spectra of the patient’s urine (Table 2 and Figure 2C). 98 Aminoacylase 1 deficiency Figure 1 One-dimensional 1H-NMR spectra (500 MHz) of urine measured at pH 2.5. A. Urine of a patient with aminoacylase 1 deficiency with assignments as marked (Table II). B. Urine of patient with aminoacylase 2 deficiency (Canavan disease). The concentration of NAA was determined to be 1470 μmol/mmol Creatinine. 99 Chapter 2.3 Figure 2 Two-dimensional NMR spectra of the urine from a patient with aminoacylase 1 deficiency. A. and B. Two regions from an 1H-13C HSQC spectrum. The 1H/13C shift values for the correlation peaks of the N- acetylated amino acids are listed in Table II C. and D. Two regions from an 1H-1H COSY spectrum. Thus, 1H and 13C chemical shifts are fully consistent with the assignment of Nacetylalanine. Similarly, the unknown triplets at 2.49 and 2.38 ppm were assigned to N-acetylglutamic acid and N-acetylglutamine, respectively. Furthermore, the 2dimensional spectra confirmed the presence of the metabolites Nacetylmethionine, N-acetylglycine and N-acetylasparagine (Fig. 2A-D). For some metabolites the singlet resonance from the CH3 protons of the N-acetyl group was the only resonance from the metabolite in the 1-dimensional spectrum. In the urine of some patients the β CH2 resonances from N-acetylasparagine could be observed. The identification of N-acetylated forms of the branched chain amino acids leucine, isoleucine and valine was based solely on the N-acetyl CH3 proton resonances in the spiking experiments. GC-MS data confirm the presence of these metabolites in clearly increased amounts in the urine of all patients [182]. In addition to the assigned singlet resonances, three smaller singlet resonances at 2.051, 2.075 and 2.108 ppm were observed in the 1H-NMR spectra of some patients. The singlet at 2.051 probably derives from the N-acetyl protons of Nacetylneuraminic acid (reference values for urine, < 20 µmol/mmol creatinine), which is always present in urine samples. Its concentration was within the reference range for age in all patients. The singlets 2.075 and 2.108 could not be assigned. These peaks were observed in all the urine samples of all 8 of our patients. Therefore, these resonances probably derive from as yet unidentified metabolites that must relate to the primary defect. Assuming three contributing 100 Aminoacylase 1 deficiency protons the concentration of these metabolites is approximately 40 µmol/mmol creatinine. To obtain more information about the chemical shift reproducibility of the Nacetyl singlets we measured the chemical shift of these peaks in the 14 urine samples from all patients at pH 2.50. All peaks show a minimal variation in the chemical shift (< 0.005 ppm; Table 2), which allows reliable identification of the metabolites. QUANTIFICATION Urine. In all patients, urine concentrations of the N-acetylated amino acids could be determined from the 1-dimensional 1H-NMR spectra. Quantification of Nacetylglutamine and of N-acetyl forms of the branched chain amino acids, leucine, isoleucine and valine, was difficult in some cases due to their rather low concentration and due to overlap with other resonances. For example, in the 1HNMR urine spectrum of patient 7 the singlet of N-acetylleucine overlapped with the singlet of N-acetylasparagine. In these cases the concentration could not be calculated. Therefore, semi quantitative estimates are given for these metabolites in table 1. Quantification of N-acetylmethionine was performed on the S-CH3 singlet (Figure 1; peak 12). In almost all cases, N-acetylglutamic acid had the highest concentration in urine varying between 170 and 740 µmol/mmol creatinine (reference value < 2 µmol/mmol creatinine [185]). Exceptions are the urine samples from patients 5 and 8, where N-acetylglycine or N-acetylasparagine were the most prominent metabolites. In patient 8 this was found consistently in three independent samples. N-Acetylmethionine resonances were observed in the urine NMR spectra of all patients. Exceptions were two urine samples from patient 7 (0.5 and 0.7 years) that had been stored at –20ºC for six months. GC-MS analysis had shown clearly increased concentrations of N-acetylmethionine in these urine samples immediately after collection. To test the hypothesis that N-acetylmethionine may disappear from urine sample during storage, the NMR measurement of the urine sample from patient 1 was repeated. Storage for 8 months at –20ºC resulted in a decrease of N-acetylmethionine from 120 to 10 µmol/mmol creatinine. All other N-acetylated amino acids showed the same concentrations in the two measurements. The low excretion of N-acetylserine and N-acetylthreonine in patient 3 (< 10 µmol/mmol creatinine) suggests that also these metabolites may have disappeared due to storage artifacts. Therefore we measured a second urine from patient 3. In this sample, the concentration of N-acetylserine and N- acetylthreonine was in the same range as in the other patients with ACY1 deficiency (Table 1). 101 Chapter 2.3 Table 2. NMR assignmentsa (urine, pH 2.5) for N-acetylated amino acids characteristic of aminoacylase 1 deficiency Metabolite Assignment δ 1H N-Acetylalanine N-Acetyl CH3 β CH3 α CH 1.41 d N-Acetylleucine N-Acetylasparagine N-Acetylglutamic acid N-Acetylglutamine N-Acetylmethionine N-Acetyl CH3 δ CH3 δ’ CH3 γ CH β CH2 α CH N-Acetyl CH3 β+β’ CH2 α CH N-Acetyl CH3 γ CH2 β+β’ CH2 α CH N-Acetyl CH3 γ CH2 β+β’ CH2 α CH N-Acetyl CH3 γ +γ’’ CH2 β+β’ CH2 α CH S-CH3 N-Acetylisoleucine N-Acetyl CH3 δ CH3 β’ CH3 γ CH2 β CH α CH N-Acetylglycine N-Acetyl CH3 α CH2 N-Acetylvalined N-Acetyl CH3 γ+γ’ CH3 β CH α CH N-Acetylserine N-Acetylthreonine N-Acetyl CH3 β+β’ CH2 α CH N-Acetyl CH3 γ CH3 β CH α CH δ 13C Peak no.c 2.018 s (2.015-2.020)b 24.6 1 4.31 m 19.2 51.9 15 2.029 s (2.027-2.031) 24.6 2 0.93 d 25.1 0.88 d 1.65 m 23.5 27.4 1.65 m 42.4 2.031 s (2.029-2.033) 24.7 4.33 m 54.9 2.83 AB 39.7 2.034 s (2.032-2.036) 24.6 2.00 m, 2.20 m 28.9 4.68 m 2.49 t 4.39 m ? 33.1 24.6 1.99 m, 2.20 m 29.5 34.3 4.36 m 55.4 2.040 s (2.038-2.042) 24.7 2.00 m, 2.15 m 33.0 2.58 m, 2.64 m 32.4 4.49 m 55.0 2.106 s (2.105-2.106) 5 13 6 17.1 12 2.043 s (2.042-2.046) 24.6 7 0.92 d 17.9 0.88 t 13.6 1.45 m, 1.24 m 27.6 4.23 m 60.9 1.91 m 39.2 2.050 s (2.048-2.051) 24.6 8 2.064 s (2.062-2.066) 24.6 9 2.17 m 32.8 3.95 d 0.95 t 4.21 dd 44.4 20.1 61.8 2.068 s (2.067-2.070) 24.7 4.48 m 58.2 2.099 s (2.097-2.100) 24.7 4.37 m 70.3 3.92 AB 1.21 d 4.42 m 64.3 21.9 61.4 a: Chemical shifts in ppm relative to TSP; multiplicities are: s, singlet; d, doublet; q, quartet, qt, quintet, m, multiplet; AB, AB system (second-order coupling) b: Variation in the chemical shift c: Numbers refer to the peaks observed in the urine 1-dimensional 1H-NMR spectrum (Fig. 1A) 102 4 14 55.3 2.039 s (2.036-2.040) 2.38 t 3 10 11 Aminoacylase 1 deficiency Serum. From six patients (patient 2-7), we analyzed serum samples using 1dimensional 1H-NMR spectroscopy. Proline also has resonances in the N-acetyl region (multiplet deriving from the ß-CH2 protons). Due to its high concentration in serum (40-332 µmol/L [110]) and the high degree of splitting identification and quantification of other metabolites in this region is hampered. We have quantified the highest peak between 2.0-2.1 ppm in the serum samples, assuming N-acetyl protons cause this peak. Excluded from this quantification was the singlet for acetic acid at 2.08 ppm. The maximal concentration was found to be 20 µmol/L. This implies that the concentration of individual N-acetylated amino acids in serum of the patients must have been lower than 20 µmol/L. Cerebrospinal fluid. We also analyzed a CSF sample from patient 7. The N-acetyl region in the 1H-NMR spectrum in control CSF samples shows several singlet resonances. In an earlier study [3], these singlet resonances were reported as common but unidentified resonances in 40 CSF samples (unknown peaks x12 [singlet 2.07], x11 [singlet 2.04 ppm] and x10 [singlet at 2.02 ppm]). These singlets may derive from N-acetylserine (2.07 ppm), N-acetylaspartylglutamic acid (2.04 ppm), N-acetylglutamic acid (2.03 ppm) and N-acetylalanine (2.02). Spiking experiments with pure N-acetylated compounds were inline with these assignments. We established reference values from 10 normal CSF samples. These amounted to <1 – 15 µmol/l. The concentration of N-acetylaspartylglutamic acid in normal CSF is comparable with the observation of Wolf et al. [25]. In CSF of patient 7 with aminoacylase 1 deficiency we found normal concentrations for all eleven N-acetylated amino acids that accumulate in this disease. N-Acetylserine, N-acetylaspartylglutamic acid, N-acetylglutamic acid and N-acetylalanine amounted to 10, 4, 6 and 4 µmol/l, respectively. AMINOACYLASE 2 DEFICIENCY (CANAVAN DISEASE) (OMIM: 271900) The urine from an affected patient was measured by 1-dimensional 1H-NMR spectroscopy and compared with the urine spectrum from a patient with ACY1 deficiency (Fig. 1A and B). The urine spectrum of the CD patient (Fig. 1B) showed the presence of high concentrations of NAA (1470 µmol/mmol creatinine; Reference: 6–36 µmol/mmol creatinine). The singlet at 2.03 ppm (N-acetyl protons) and the characteristic β CH2 proton resonances at 2.95 ppm confirm the presence of NAA in the urine of the patient. Figure 1 shows that NMR spectroscopy can easily discriminate between ACY1 and ACY2 deficiency. 103 Chapter 2.3 OTHER N-ACETYLATED AMINO ACIDS In order to improve assignment in the N-acetyl region of body fluid NMR spectra other N-acetylated amino acids were measured. Table 3 provides on overview of fourteen N-acetylated amino acids. Some of these metabolites assist in the diagnosis of various inborn errors of metabolism as indicated in Table 3. Table 3. 1H chemical shiftsa (pH 2.5) for N-acetylated amino acids observed in urine of patients with various inborn errors of metabolism and other relevant to human metabolism. Metabolite Assignment δ 1H N-Acetylarginine N-Acetyl CH3 δ CH2 γ CH2 β+β’ CH2 α CH 2.04 s N-Acetyl CH3 β+β’ CH2 α CH 2.03 s N-Acetyl CH3 Aspartyl: β+β’ CH2 α CH Glutamic acid: γ+γ” CH2 β+β’ CH2 α CH 2.04 s - Canavan disease 2.85 AB - Severe hypomyelination associated with increased levels N-Acetyl CH3 δ CH2 γ CH2 β+β’ CH2 α CH 2.03 s N-Acetyl CH3 β+β’ CH2 α CH 2.07 s N-Acetyl CH3 β+β’ CH2 α CH 1.99 s C2H, ring 8.59 s N-Acetylaspartic acid N-Acetylaspartylglutamic acid N-Acetylcitrulline N-Acetylcysteine N-Acetylhistidine C4H, ring N-Acetylisoleucine N-Acetylleucine N2-Acetyllysine 104 See table 2 See table 2 N-Acetyl CH3 ε CH2 δ CH2 γ CH2 β+β’ CH2 α CH Inborn errors of metabolism 3.21 q 1.66 m 1.91 m, 1.77 m 4.32 m - Canavan disease (Figure 1B) 2.95 m 4.72 m 4.72 dd - Pelizeus Merzbacher Disease of N-acetylaspartyl-glutamate in CSF[25] 2.01 m, 2.23 m 2.48 m 4.44 dd 3.12 t - Citrullinemia 1.55 m 1.86 m, 1.73 m 4.30 m 2.97 m 4.57 m 3.15m - Histidinaemia 4.67m 7.29 s - Maple Syrup Urine Disease 2.03 s 3.00 m 1.67 m 1.44 m 1.88 m, 1.75 m 4.30 m - Maple Syrup Urine Disease - Hyperlysinaemia - Lysinuria Aminoacylase 1 deficiency Table 3. 1H chemical shiftsa (pH 2.5) for N-acetylated amino acids observed in urine of patients with various inborn errors of metabolism and other relevant to human metabolism. (Continued) N5-Acetyllysine N-Acetylornithine N-Acetylphenylalanine N-Acetylproline N-Acetyltaurine N-Acetyltryptophan N-Acetyltyrosine N-Acetylvaline N-Acetyl CH3 ε CH2 δ CH2 γ CH2 β+β’ CH2 α CH 1.97 s N-Acetyl CH3 δ CH2 γ +γ’ CH2 β CH2 α CH 2.04 s N-Acetyl CH3 β+β’ CH2 α CH 1.93 s H, ring 7.33 various 3.17 q 1.55 m 1.41 m, 1.89 m 1.41 m, 1.89 m 3.84 t 3.05 1.01 1.94 4.35 m - Phenylketonuria (PKU) 3.12 m 4.61 m N-Acetyl CH3 δ CH2 γ CH2 β+β’ CH2 α CH 2.11 s N-Acetyl CH3 1.99 s S-CH2 3.55 m 3.42 m 2.00 m 2.00 m 4.40 dd N-CH2 3.08 t N-Acetyl CH3 β+β’ CH2 α CH 1.92 s H, ring 7.2 – 7.7 Various - Liver diseases 3.31 m 4.70 qu N-Acetyl CH3 β+β’ CH2 α CH 1.94 s - Tyrosinemia type I 4.57 m - Liver diseases H, ring 7.89 AB See table 2 3.07 AB - Tyrosinemia type II - Maple Syrup Urine Disease a: Chemical shifts in ppm relative to TSP; multiplicities are: s, singlet; d, doublet; q, quartet, qt, quintet, m, multiplet DISCUSSION This study describes the in vitro NMR spectroscopic investigations at the metabolite level on eight unrelated patients with ACY1 deficiency. The 1dimensional 1H-NMR urine spectra from these patients show a characteristic metabolic profile, mainly due to the N-acetyl singlets deriving from N-acetylated amino acids. Using a combination of one dimensional 1H and two dimensional 1H1H COSY and 1H-13C HSQC NMR experiments, resonances of 11 N-acetylated amino acids could be assigned and quantified or semi-quantified. The profile of 105 Chapter 2.3 the accumulating N-acetylated amino acids is consistent with the enzyme specificity of ACY1. The enzyme preferentially uses neutral aliphatic amino acids with a short-chain acyl moiety as substrate [178]. These include the N-acetylated amino acids of serine, threonine, alanine, methionine and glycine. N-Acetylated branched chain amino acids can also be a substrate for ACY1 [178]. This explains the presence of N-acetylated forms of leucine, valine and isoleucine in the urine of our patients. The same holds for the N-acetylated forms of the basic amino acids glutamine and asparagine and the acidic amino acid glutamic acid. NAcetylaspartic acid (=NAA) is not a substrate for ACY1 [178]. NAA is one of the most abundant amino acid in the human brain. There, it is catabolized by a special aminoacylase (ACY2) to acetate and aspartate. NAA therefore does not accumulate in body fluids of the patients with ACY1 deficiency. N-Acetylated aromatic amino acids are not good substrates for ACY1 [178]. Aminoacylase 3 and acyl-lysine deacylase (EC 3.5.1.17) are two other enzymes involved in the breakdown of Nacetylated amino acids [178]. Aminoacylase 3 acts preferentially on N-acylaromatic amino acids, e.g. N-acetyltryptophan, N-acetyltyrosine, Nacetylphenylalanine and N-acetylhistidine and the enzyme acyl-lysine deacylase catalyzes the N-deacetylation of N5-acetyllysine. Accordingly, none of these Nacetylated amino acids were observed in increased concentration in the urine of the ACY1 deficient patients. A deficiency of aminoacylase 3 has never been described. It is to be expected that body fluid NMR spectroscopy would easily pick up the resonances of N-acetylated aromatic amino acids and therefore may be a good technique to diagnose patients with aminoacylase 3 deficiency. Also Nacetylated aromatic amino acids will be picked up in GC-MS analysis of organic acids [182]. Differences in excretion levels were observed between various N-acetylated amino acids in the urine of our patients. Table 1 shows that the N-acetylated form of glutamic acid had the highest concentration in all patients except patients 5 and 8 where N-acetylglycine or N-acetylasparagine were the prominent diagnostic metabolites. Such variations may relate to the specific mutations of a patient but higher number of patients would be required to prove this hypotheses. In general, the excretion levels of the N-acetylated amino acids do not correlate with serum concentrations of the corresponding free amino acid. Most of the accumulating Nacetylated amino acids are excreted in a concentration range between 100-500 µmol/mmol creatinine. Exceptions are the N-acetylated forms of the branched chain amino acids and threonine and methionine. The excretion of Nacetylthreonine for instance ranges between 20 and 60 µmol/mmol creatinine. It may be that these N-acetylated amino acids are accepted as a substrate by one of the other aminoacylases. In addition, the concentration of the N-acetylated amino acids in our patients may partially reflect the extent of N-acetylation of specific amino acids at the N-terminus of human proteins. In mammals N-terminal acetylation of amino acids occurs on approximately 80-90% of proteins [186]. 106 Aminoacylase 1 deficiency Severe neurological involvement has been observed in individual patients with ACY1 deficiency [54;55]. However, CSF of only 1 patient was available for analysis and therefore the observations on this sample are preliminary. We found normal concentrations of N-acetylated amino acids in CSF of patient 7. Further measurements of CSF of affected cases are required to confirm our observation. Our observation may reflect that ACY1 deficiency does not lead to accumulation of N-acetylated amino acids in the brain. It will be of interest to see whether accumulation of N-acetylated amino acids can be demonstrated in the brain of these patients with in vivo MRS. ACY1 gene expression has been found in brain tissue and ACY1 protein and enzyme activity have been demonstrated in brain [54;178]. Normal levels of N-acetylated amino acids in CSF in patients with ACY1 deficiency question the role of the enzyme in brain tissue. In ACY2 deficiency (Canavan disease) NAA does accumulate in CSF (380 µmol/l; reference value: < 2 µmol/l [3;187]). The relation between the molecular defect and the neurological signs and symptoms is obvious in Canavan disease. In ACY1 deficiency, this is less obvious even more so when we take into account the asymptomatic case described by Sass et al [54], patient 3 in this study. However, this patient still is at very young age and may develop clinical signs and symptoms later in life. Initially, the high amounts of N-acetylated amino acids in the urine of patients 1 – 8 were detected by routine GC-MS analyses. Organic acid analysis with GC-MS is appropriate for identifying individuals with ACY1 or ACY2 deficiency. However, not all relevant N-acetylated amino acids can be detected with GC-MS in the urine of ACY1 patients. Depending on the derivatization procedure used N-acetylglutamine [54;55], N-acetylasparagine [54;55], N-acetylserine [54] or N-acetylthreonine [54] were missed by GC-MS. NMR spectroscopy does not require extensive sample treatment and using a combination of one- and two-dimensional NMR spectra, we were able to detect 11 N-acetylated amino acids in the urine of ACY1 deficient patient. Moreover, most N-acetylated amino acids could be quantified or semi quantified. In conclusion, we have shown that NMR spectroscopy can be used to identify and quantify characteristic N-acetylated amino acids in urine from patients with ACY1 deficiency. The technique can discriminate reliably between deficiencies of aminoacylase 1 and aminoacylase 2. 107 108 THREE THE COMBINATION OF IN VIVO AND IN VITRO NMR SPECTROSCOPY 109 110 CHAPTER 3.1 Dimethyl sulfone in human cerebrospinal fluid and blood plasma confirmed by one-dimensional 1H and two-dimensional 1H-13C NMR Udo F. H. Engelke, Albert Tangerman, Michèl A.A.P. Willemsen, Detlef Moskau, Sandra Loss, S. Harvey Mudd and Ron A. Wevers NMR in Biomedicine 2005; 18(5): 331-336 111 112 DMSO2 in CSF and blood plasma Abstract 1H-NMR spectroscopy at 500 MHz was used to confirm that a previously unidentified singlet resonance at 3.14 ppm in the spectra of cerebrospinal fluid and plasma samples corresponds to dimethyl sulfone (DMSO2). A triple resonance inverse cryogenic NMR probe, with pre-amplifier and the RF-coils cooled to low temperature, was used to obtain an 1H-13C HSQC spectrum of CSF containing 8 µM (753 ng/ml) DMSO2. The 1H-13C correlation signal for DMSO2 was assigned by comparison with the spectrum from an authentic reference sample. In plasma and CSF from healthy controls, the concentration of DMSO2 ranged between 0 and 25 µmol/l. The concentration of DMSO2 in plasma from three of four patients with severe methionine adenosyltransferase I/III (MAT I/III) deficiency was about twice the maximum observed for controls. Thus, DMSO2 occurs as a regular metabolite at low micromolar concentrations in cerebrospinal fluid and plasma. It derives from dietary sources, from intestinal bacterial metabolism and from human endogenous methanethiol metabolism. 113 Chapter 3.1 INTRODUCTION 1H-NMR spectroscopy of body fluids is a powerful tool in the field of inborn errors of metabolism. The technique provides an overview of proton-containing metabolites, is rapid, non-destructive and requires minimal sample treatment. For most compounds detected, quantitative analysis is possible. A prerequisite for the analysis of body fluid with NMR spectroscopy lies in the correct assignments of resonances. Several papers have reported assignments of resonances in 1H-NMR spectra of body fluids [3;4;17]. However, some resonances are still unassigned. In earlier papers we identified 16 unassigned resonances in cerebrospinal fluid (CSF) [3] and 14 in plasma [4]. In this study, we concentrate on the origin of the unidentified singlet resonance at 3.14 ppm that was described as unknown in both studies [3;4]. Here, we assign this singlet resonance to the six equivalent protons in dimethyl sulfone (DMSO2), using one-dimensional and heteronuclear two-dimensional NMR techniques. We present reference values for DMSO2 in human plasma and CSF. Furthermore, we document an increased concentration of DMSO2 in plasma from three out of four unrelated patients with methionine adenosyltransferase I/III (MAT I/III) deficiency. Our data demonstrate the common presence of DMSO2 in human blood and CSF and show that DMSO2 in body fluids derives partly from dietary sources, partly from bacterial metabolism and partly from endogenous human methanethiol metabolism. EXPERIMENTAL CONTROL SUBJECTS CSF samples were obtained from 30 patients (age range 0–78 years) suspected to suffer from neurometabolic disease. The clinical samples were sent to our laboratory for metabolic screening. Appropriate analysis ruled out immunologic and chronic infectious diseases as well as disorders caused by toxic agents. For these subjects normal results were obtained for metabolic screening tests of urine (amino acids, organic acids and lysosomal diseases), and the CSF samples were classified as ‘controls’. Plasma samples were obtained from 15 healthy adult volunteers (age range 24–66 years; seven male and eight female), who gave their informed consent. 114 DMSO2 in CSF and blood plasma PATIENTS WITH MAT I/III DEFICIENCY Plasma samples were available from four severely and two mildly affected MAT I/III-deficient patients (OMIM 250850). All six patients have mutations in the gene MAT1A on chromosome 10q224 [188] that encodes the catalytic subunit of the two native MAT isoenzymes (EC 2.5.1.6), MAT I and MAT III [189]. SAMPLE PREPARATION Plasma and CSF samples were deproteinized using a 10 kDa filter (Sartorius). Before use, the filter was washed twice by centrifugation of water to remove glycerol. A 20 µl aliquot of 20.2mM trimethylsilyl-2,2,3,3- tetradeuteropropionic acid (TSP, sodium salt; Aldrich) in 2H2O was added to 700 µl of the ultrafiltrate, providing a chemical shift reference (δ=0.00), a concentration reference and a deuterium lock signal. The pH of the ultrafiltrate was adjusted to 2.50±0.05 with concentrated HCl. Finally, 650 µl of the sample was placed in a 5mm NMR tube (Wilmad Royal Imperial). ONE-DIMENSIONAL 1H-NMR SPECTROSCOPY Plasma and CSF were measured at 500MHz on a Bruker DRX 500 spectrometer equipped with a triple-resonance inverse (TXI) 1H{15N, 13C} probehead and equipped with x, y, z gradient coils. 1H spectra were acquired as 128 transients in 32K data points with a spectral width of 6002 Hz. The sample temperature was 298K and the H2O resonance was presaturated by single-frequency irradiation during a relaxation delay of 10 s, and a 90º excitation was used. Shimming of the sample was performed automatically on the deuterium signal. To improve the spectral resolution, the samples were spun (7 Hz) during the measurements. The resonance linewidths for TSP and metabolites were <1 Hz. A π/2-shifted sine-bell window function was applied to the FID. Fourier transformation was performed after zero-filling to 64K data points. The phase and the baseline were corrected manually. The DMSO2 singlet resonance (six equivalent protons from the two methyl groups) and the TSP singlet (nine equivalent protons) were fitted semiautomatically with Lorentzian line shapes. The concentration of DMSO2 was calculatedfrom the relative integrals of the fitted lineshapes using the known concentration of TSP. 1H-13C HETERONUCLEAR SINGLE-QUANTUM CORRELATION (HSQC) SPECTROSCOPY An HSQC spectrum of one CSF sample was measured at 500MHz on a Bruker AV 500 spectrometer using an inverse triple-resonance cryoprobe (Bruker), whereby the preamplifier and the RF coils were cooled to low temperature. The sequence was optimized for a 1J CH-coupling constant of 130 Hz. FIDs (1536 transients) were acquired with 2K time-domain data points (t2), an acquisition time of 0.128s, a 115 Chapter 3.1 recycle time of 2.128s and 96 t1 increments. The spectral widths were 16 ppm (1H) and 70 ppm (13C), and the total measurement time was 5 days and 10 h. RESULTS LINEARITY, RECOVERY AND DETECTION LIMIT To study linearity and recovery we added DMSO2 in known concentrations (0, 32, 62, 94 and 195 µmol/l) to a pool of control CSF samples. The recovery was 90– 100% at all concentration levels. The linearity and correlation coefficient were excellent (95% confidence interval of the slope, 0.882–0.996; correlation coefficient, r>0.99). The detection limit at S/N=3 for DMSO2 was determined to be 0.5 µmol/l (1H-NMR at 500 MHz, 128 transients; using the conventional 5mm TXI probehead operating at ambient temperature). ONE-DIMENSIONAL 1H-NMR SPECTROSCOPY An expanded region (2.50–3.20 ppm) of the 1H-NMR spectrum of a control CSF sample is shown in Figure 1 for pH 7.5 (lower panel) and pH 2.5 (upper panel). The major metabolites in this region and the chemical shifts at pH 2.5 are creatinine (singlet, 3.13 ppm), creatine (singlet, 3.05 ppm) and citric acid ([AB]2 spin system, 2.99, 2.83 ppm). The singlet resonance at 3.14 ppm was observed in an earlier CSF study as unidentified resonance x14 and was assigned to DMSO2 as described below [3]. Figure 1 1H-NMR spectra (500 MHz) of a representative CSF sample measured at pH 2.5 (upper panel) and pH 7.5 (lower panel). The concentration of dimethyl sulfone was determined to be 18 µmol/l 116 DMSO2 in CSF and blood plasma IDENTIFICATION OF DMSO2 NMR spectroscopy was performed on CSF from a 3-year-old boy with mental retardation and loss of acquired skills. Routine metabolic screening of the patient’s body fluids showed no abnormalities. However, an unusually high concentration of the metabolite corresponding to the unknown singlet x14 was observed in the one-dimensional 1H-NMR CSF spectrum (data not shown) [3]. This observation could not be confirmed with a repeat CSF sample. It was then discovered that at the time of the first lumbar puncture this patient had used a dietary supplement containing DMSO2. 1H-NMR of the authentic compound DMSO2 in H2O at pH 2.5 showed a singlet at 3.14 ppm. The addition of pure DMSO2 to a control CSF sample (spiking) resulted in an increase of the singlet at 3.14 ppm. Additionally, one-dimensional 1H-NMR spectra of a CSF sample and reference solution of DMSO2 in H2O were recorded at pH 2.5 and 7.5. Both samples exhibited a singlet resonance which shifted from 3.138 ppm at pH 2.5 to 3.157 ppm at pH 7.5 (Figure 1). These data provide strong evidence for the assignment of DMSO2 as source of the singlet x14. Furthermore, at pH 7.5, characteristic shifts were observed for the creatinine peak (3.13 at pH 2.5 to 3.05 ppm at pH 7.5) and citric acid [AB]2 system (2.99, 2.83 ppm at pH 2.5–2.70; 2.54 ppm at pH 7.5). CONFIRMATION OF DMSO2 BY 1H/13C HSQC TWO-DIMENSIONAL NMR SPECTROSCOPY Additional confirmation of the DMSO2 in body fluid was provided by 1H-13C HSQC two-dimensional NMR experiments performed on an authentic reference solution of DMSO2 and a control CSF sample containing 8 µmol/l DMSO2. The twodimensional spectrum of the reference sample exhibited a correlation peak with the coordinates 3.14 (1H)/44.3 (13C) (data not shown), and a resonance was observed at the same position in the two-dimensional spectrum of control CSF sample in addition to correlation signals for the other metabolites shown in Figure 1. Thus, 1H and 13C chemical shifts and the pH dependence of the 1H shift are fully consistent with the assignment of the previously unknown metabolite as DMSO2. OCCURRENCE AND CONCENTRATION OF DMSO2 IN CSF AND PLASMA We observed the DMSO2 singlet resonance in 28 of 30 clinical CSF samples examined, consistent with the results reported earlier (35 out of 40 samples) [3]. Furthermore, DMSO2 was detected in 11 of 15 plasma samples from a different group of healthy volunteers. No age and sex dependency for DMSO2 was found in blood and CSF. Figure 2 summarizes DMSO2 concentration determined for individual uncorrelated plasma and CSF samples. The highest concentration amounted to 25 µmol/l in CSF (mean±SD, 11.3±6.2) and 24 µmol/l in plasma 117 Chapter 3.1 (mean±SD, 8.8±7.3). A much higher concentration of DMSO2 in plasma and CSF was observed for the one patient receiving a dietary supplement containing DMSO2 (plasma, 624 µmol/l; CSF, 574 µmol/l). We also measured the concentration of DMSO2 in plasma CSF sample pairs obtained within 1 h from four control patients not included in the groups presented in Figure 2. In all four pairs, the concentration of DMSO2 was slightly higher in plasma (plasma/CSF values, each in µmol/l: 7:2; 15:8; 13:11; 20:15). Correlation analysis (Fischer’s r to z) gave a correlation coefficient r=0.928 (p=0.10, n=4). Figure 2 Quantification of DMSO2 in plasma and CSF from adults. ({) Plasma from healthy adult volunteers (n=15); ( ) CSF samples classified as ‘controls’ (n=30); () plasma from severely affected patients with MAT-deficiency (methionine>300 µmol/l; n=4); (z) plasma from mildly affected patients with MAT-deficiency (methionine< 300 mmol/l; n=2). Values of zero correspond to DMSO2 levels below the detection threshold of 0.5 µmol/l DMSO2 IN PLASMA FROM MAT I/III-DEFICIENT PATIENTS MAT catalyzes the biosynthesis of S-adenosylmethionine (AdoMet) from methionine. Patients with the disease have high concentrations of methionine and methionine sulfoxide in urine and plasma. The normal trans-sulfuration pathway is defective in these patients and methionine degradation via the transamination pathway is increased [190] (Figure 3). Methionine transamination metabolites accumulate abnormally only when plasma methionine concentrations exceed 300– 350 µM [191]. Dimethyl sulfide (DMS) is such an intermediate in this pathway. We measured the DMSO2 concentration in plasma from different unrelated patients with MAT I/III deficiency. In four severe patients, the methionine level in plasma was higher than 300 µmol/l. Here, the concentration of DMSO2 was significantly higher than for the 15 controls (p=0.04, one-sided Wilcoxon test; Figure 2). For one patient with severe deficiency and two patients with mild deficiency (methionine <300 µmol/l), the plasma concentration of DMSO2 was within the control range. Some of the samples from MAT I/III-deficient patients were stored 118 DMSO2 in CSF and blood plasma at -70ºC for several years so that elevated DMSO2 concentrations could be the result of in vitro oxidation of DMS during storage. To test this hypothesis, DMS was added to plasma at concentrations of 30 and 60 µmol/l in a closed 15 ml septum vial. Storage for one week at room temperature did not result in the formation of detectable DMSO2 and the added DMS was recovered quantitatively (data not shown). Thus, in vitro oxidation of DMS is unlikely to be a source of DMSO2. The one-dimensional 1H spectra from the plasma obtained from the six patients all showed the presence of the methionine singlet at 2.13 ppm and the methionine sulfoxide singlet at 2.74 ppm (normally not present). Dimethyl sulfoxide (DMSO) (singlet at 2.72 ppm) was not detectable in patient plasma. Figure 3 Transamination pathway for methionine and metabolic steps leading to dimethyl sulfone. Methionine adenosyltransferase I/III deficiency (OMIM 250850) results in a strong reduction (bar) in methionine flux through the trans-sulfuration pathway. 119 Chapter 3.1 DISCUSSION The results of this study confirm the presence of DMSO2 as a common metabolite in human CSF and plasma. The higher detection sensitivity achievable with CryoProbe technology allowed the 1H-13C signal of 8 µmol/l DMSO2 to be assigned in an HSQC spectrum of a control CSF sample. As early as 1959 Williams [192] suggested that a possible fate of DMS would be oxidation to DMSO2. To our knowledge Williams and Burstein were the first and only authors to describe the presence of DMSO2 in urine. They reported concentrations of 4–11 mg/day (~20– 120 µmol/l) [193] in urine of healthy volunteers. Jacob et al. [194] mentioned DMSO2 in blood of an adult man (2 µM). Until now, DMSO2 had not been described as a regular metabolite in CSF. Commodari et al. [20] described a singlet resonance at 3.17 ppm at pH 9.6 in two out of eight CSF samples (peak Y) and Wevers et al. [3;4] reported this singlet at 3.14 ppm (pH 2.50) in plasma and CSF spectra (peak u10 in plasma and peak x14 in CSF). However, in none of these in vitro NMR studies could this singlet be definitively assigned to DMSO2. Our data for the patient receiving DMSO2 in a dietary supplement suggest that intestinal DMSO2 can enter the blood and subsequently reach the CSF. Moreover, threepapers have reported DMSO2 in the human brain, as detected by in vivo 1HMRS [195-197]. In all cases the presence of this metabolite resulted from ingestion of a dietary supplement containing DMSO2. Milk [198] and port wine [199] are also possible sources of DMSO2, and we confirmed its presence in milk (20 µmol/l). Until now, it was unknown if other food sources or beverages contained DMSO2. Another exogenous source of DMSO2 is the sulfur oxidation of dietary DMS via DMSO to DMSO2 [200]. DMS has been found in cooked vegetables [193], beer (0.2–2.3 µmol/ l) [201], rice (21–319 µmol/kg) [202] and port wine (up to 0.8 µmol/ l) [199;203]. However, the relatively low concentrations of DMSO2 and DMS found in food and beverages cannot explain the concentrations of DMSO2 found in CSF and plasma. Therefore, it appears likely that DMSO2 can also be a product of endogenous metabolism. In the gut methanethiol is formed either from methionine or from methylation of hydrogen sulfide, and subsequent methylation of methanethiol results in the formation of DMS [204]. It has been shown that methanethiol can be converted to DMS. Furne et al. [205] have postulated that conversion of methanethiol to thiosulfate and sulfate (via formation of H2S) is much more predominant in rat intestine than conversion of methanethiol to DMS. Some of the methanethiol directly enters the blood [206]. Once in the blood, methanethiol is nearly quantitatively oxidized by erythrocytes to formic acid and sulfate (ca. 99%; Figure 4). Only 1% of the total methanethiol is metabolized to DMS by the enzyme thiol methyltransferase [206]. DMS may then be further oxidized to DMSO and DMSO2. Methanethiol, DMS, DMSO and DMSO2 all seem able to enter the blood. 120 DMSO2 in CSF and blood plasma Subcutaneous injection of DMS leads to the excretion of DMSO and DMSO2 in the urine of rabbits [200]. Rhesus monkeys treated with DMSO had high concentrations of DMSO2 in their blood [207]. The authors concluded that part of the DMSO was also metabolized to DMS, as illustrated by a particular sweetness of the breath of the monkeys. These data are strongly indicative for both intestinal bacterial metabolism of methanthiol and also from endogenous human methanethiol metabolism as possible sources for body fluid DMSO2. Human endogenous methanethiol conversion into DMSO2 is further supported by our finding of an increased plasma DMSO2 concentration in patients with MAT I/III deficiency. In these patients the normal trans-sulfuration pathway of methionine is defective and the transamination of methionine occurs more extensively (Figure 3), especially when plasma methionine concentrations exceed 300 µM. These patients may have high levels of DMS, a product of the transamination pathway, in their breath [191]. In three out of four plasma samples from severely affected patients with MAT I/III deficiency, the DMSO2 plasma concentration was about twice the maximum value for controls. Figure 4 Both dietary sources and putative endogenous metabolic pathways can potentially contribute to the presence of dimethyl sulfone in human body fluids 121 Chapter 3.1 This elevation is most likely the result of oxidation of elevated levels of DMS via DMSO to DMSO2 (Figure 4) and indicates that at least some fraction of body fluid DMSO2 may derive from methionine metabolism. In one severely affected patient with MAT I/III deficiency, the DMSO2 concentration fell within the normal range. An explanation for this phenomenon might be that in this patient the methylation of methanethiol to dimethyl sulfide is partly blocked, resulting in normal DMSO2 values. The patient receiving a dietary DMSO2 supplement exhibited high concentration of DMSO2 in both CSF and plasma (ca. 600 µmol/l). The plasma: CSF DMSO2 ratio of ca. 1.09 suggests rapid equilibration of this metabolite between blood and CSF compartments. Our results are consistent with other MRS studies that show that dietary DMSO2 accumulates in the brain [195-197]. Plasma and CSF studies using dogs showed high brain concentrations of DMSO after in vivo administration [208], indicating that the transfer of DMSO from blood into CSF is also possible. In conclusion, we have shown that DMSO2 is a common metabolite in human CSF and plasma, occurring at concentrations of 0–25 µmol/l. Furthermore, we have shown that the DMSO2 concentration may be higher in the plasma of severely affected MAT I/III-deficient patients, and that the DMSO2 detected in body fluids probably derives from both dietary sources and endogenous metabolism. 122 CHAPTER 3.2 Methylsulfonylmethane (MSM) ingestion causes a significant resonance in proton magnetic resonance spectra of brain and cerebrospinal fluid Michèl A.A.P. Willemsen, Udo F.H. Engelke, Marinette van der Graaf and Ron A. Wevers Neuropediatrics 2006; 37(5): 312-314 123 124 MSM in brain Abstract The use of methylsulfonylmethane (MSM), available as an ‘over-the-counter’ dietary supplement, led to the occurrence of an abnormal resonance at 3.15 ppm in the in vivo brain proton MR spectrum as well as the in vitro cerebrospinal fluid NMR study of a 4-year-old girl. The concentration of this compound amounted to 1.2 mmol/l in brain tissue and 1.7 mmol/l in cerebrospinal fluid. Our findings illustrate that ingestion of exogenous compounds, e.g. in medication, food or ‘innocent’ supplements, may lead to abnormal resonances in spectroscopy studies that might be difficult to assign. 125 Chapter 3.2 Patient A 4-year-old girl and her 6-year-old brother both suffered from a progressive demyelinating sensomotor polyneuropathy that started at the age of 1½ years and 6 months respectively. The neurological disorder in the boy additionally included mild mental retardation and mild bilateral pyramidal tract involvement. Extensive work-up had never led to a classifying diagnosis in these siblings. The girl underwent routine magnetic resonance (MR) imaging and proton MR spectroscopy investigations in a 3T Magnetom Trio MR System (Siemens, Erlangen, Germany) using a standard CP head coil. Single voxel MR spectroscopy was performed using PRESS volume selection (TE/TR=20/5000 ms) with and without water suppression. After this procedure a lumbar puncture was performed. Nuclear magnetic resonance (NMR) spectroscopy was performed on CSF using methods described previously [10]. Initially the parents had reported that the girl did not use any medication at that time, but after the investigations (see below) they told us that the girl had methylsulfonylmethane as a dietary supplement. Figure 1 3T single voxel proton MR spectrum (PRESS, TE/TR=20/5000 ms) of occipital gray matter (7.2 ml) clearly showing a resonance of methylsulfonylmethane (MSM) at 3.15 ppm between the methyl signals of choline (3.2 ppm) and creatine (3.0 ppm). The voxel location is shown in the MR image. 126 MSM in brain Results MR imaging showed normal cerebral anatomy and myelination. Proton MR spectroscopy of cerebral white matter, basal ganglia as well as occipital gray matter (Figure 1) showed an abnormal resonance at 3.15 ppm in an otherwise normal spectrum. The abnormal resonance could be attributed to the presence of methylsulfonylmethane (MSM; synonyms: dimethyl sulfone, sulfonylbismethane). An approximation of the amount of MSM in occipital gray and periventricular white matter was determined by comparison of the signal area of the resonance of dimethyl sulfone at 3.15 ppm and the methyl signals of creatine (3.03 ppm) and choline (3.22 ppm). Assuming comparable attenuation of the three resonances due to relaxation and assuming choline and creatine contents of 0.88 mM and 6.9 mM respectively, in occipital gray matter [209], the MSM concentration at this location was estimated 1.2 to 1.3 mM. A similar calculation for the voxel in periventricular white matter resulted in an MSM concentration of 1.1 to 1.3 mM. These concentrations show that the MSM is equally distributed among gray and white matter. Cerebrospinal fluid (CSF) analysis revealed normal values for cells, protein, glucose, lactate and pyruvate. Nuclear magnetic resonance (NMR) spectroscopy was performed on CSF and showed the singlet resonance of MSM at 3.14 ppm (at pH 2.5; Figure 2). The concentration of MSM in CSF was 1.7 mmol/l. Figure 2 In vitro NMR spectrum of cerebrospinal fluid showing a resonance of dimethylsulfone (methylsulfonylmethane, MSM) at 3.14 ppm (pH 2.5). 127 Chapter 3.2 Discussion In vivo and in vitro proton MR spectroscopy (of the brain and CSF, respectively), generally referred to as 1H-MRS and body fluid NMR spectroscopy, are relatively novel techniques that are used to investigate patients who suffer from known and unknown neurological disorders. The study of peak heights of known resonances and the appearance of ‘novel’ resonances has proven to be very useful in the elucidation of pathophysiological processes in brain disorders. Sometimes, however, the interpretation of the spectral abnormalities can be difficult, especially when abnormal resonances cannot simply be assigned to known metabolites. A recent study in our own group concentrated on a hitherto unknown resonance at 3.14 ppm observed in CSF NMR spectra of some patients [10]. The resonance was assigned to methylsulfonylmethane or dimethylsulfone which appears to occur in low concentration (0 – 25 µmol/l) as a physiological metabolite in human CSF [10]. Shortly after this study we encountered the in vivo spectrum of the patient described here. The interpretation of her cerebral proton MR spectrum was therefore simple, but might otherwise have led us to search for an explanation for the abnormal resonance in this family with two children with an unknown – most probably genetic – disorder. MSM may derive from dietary sources, from intestinal bacterial metabolism and from human endogenous methanethiol metabolism [10;210]. The molecule can easily cross the blood-brain barrier. After ingestion of ‘generally advised dosages’ it can be demonstrated in millimolar concentrations in brain tissue and in CSF which is more than thousand-fold the normal physiological concentration [10;195-197;210]. Comparable results of cerebral proton MR spectroscopy have previously been described in a small number of adults [196;197] and in only one (5-year-old) child [195], all using MSM. We are surprised about the high concentrations that occur in the central nervous system after ingestion of MSM in ‘recommended’ dosages. Apparently, its accumulation is tolerated without adverse effects [10;195-197;210]. The occurrence of an abnormal resonance at 3.15 ppm on cerebral proton MR spectroscopy (or CSF NMR) should remind the investigator to check the use of MSM as dietary supplement in the patient. More in general, in vitro and in vivo MR spectroscopy might incidentally reveal puzzling results due to exogenous metabolites like medication or food additives. Other compounds that may be encounted as remarkable or “confusing” resonances in in vivo and in vitro MR spectra are e.g., propanediol (propylene glycol) and ethanol [3;211;212]. 128 CHAPTER 3.3 NMR spectroscopic studies on the late onset form of 3-methylglutaconic aciduria, type I and other defects in leucine metabolism Udo F.H. Engelke, Berry Kremer, Leo A.J. Kluijtmans, Marinette van der Graaf, Éva Morava, Ference J. Loupatty, Ronald J.A. Wanders, Detlef Moskau, Sandra Loss, Erik van den Bergh and Ron A. Wevers NMR in Biomedicine 2006; 19(2): 271-278 129 130 3-Methylglutaconic aciduria type I Abstract A diagnosis of 3-methylglutaconic aciduria type I (OMIM: 250950) based on elevated urinary excretion of 3-methylglutaconic acid (3MGA), 3-methylglutaric acid (3MG) and 3-hydroxyisovaleric acid (3HIVA) was made in a 61-year-old female patient presenting with leukoencephalopathy slowly progressing over more than 30 years. The diagnosis was confirmed at the enzymatic and molecular level. In vivo brain MR spectroscopic imaging (MRSI) was performed at 3.0 T, and onedimensional and two-dimensional in vitro NMR spectroscopy of body fluids of the patient was performed at 11.7 T. Additionally, we measured 1D 1H-NMR spectra of urine of seven patients with a total of four different inborn errors of leucine metabolism. Increased concentrations of 3HIVA, 3MGA (cis and trans) and 3MG were observed in the NMR spectra of the patient’s urine. In the cerebrospinal fluid, the 3HIVA concentration was 10 times higher than in the plasma of the patient and only the cis isomer of 3MGA was observed. In vivo brain MRSI showed an abnormal resonance at 1.28 ppm that may be caused by 3HIVA. Comparison of 1H-NMR spectra of urine samples from all eight patients studied, representing five different inborn errors of leucine metabolism, showed that each disease has typical NMR characteristics. Our leukoencephalopathy patient suffers from a late-onset form of 3-methylglutaconic aciduria type I. In the literature, only very few adult patients with this conditions have been described, and 3HIVA accumulation in white matter in the brain has not been presented before in these patients. Our data demonstrate that 1H-NMR spectroscopy of urine can easily discriminate between the known inborn errors of leucine metabolism and provide the correct diagnosis. 131 Chapter 3.3 Introduction 3-Methylglutaconic aciduria type I (OMIM: 250950) is an inborn error of leucine metabolism caused by the deficiency of the enzyme 3-methylglutaconyl-CoA hydratase (3MGH; EC 4.2.1.18). This enzyme catalyzes the conversion of 3methylglutaconyl-CoA to 3-hydroxy-3-methyl-glutaryl-CoA in leucine metabolism (step E, Figure 1). Figure 1 Metabolic pathway of leucine. 3-Methylglutaconyl CoA hydratase deficiency (solid square) results in accumulation of 3MGA, 3MG and 3HIVA. A, branched-chainamino-acid transaminase (EC 2.6.1.42); B, branched-chain 2-keto acid dehydrogenase (EC 1.2.4.4); C, isovaleryl-CoA dehydrogenase (EC 1.3.99.10); D, 3-methylcrotonyl CoA carboxylase (EC 6.4.1.4); E, 3-methylglutaconyl CoA hydratase (EC 4.2.1.18); F, 3-hydroxy-3-methylglutaryl-CoA lyase (EC 4.1.3.4); G, enoyl-CoA hydratase (EC 4.2.1.17). The metabolic phenotype of this disease is characterized by increased excretion of 3-methylglutaconic acid (3MGA), 3-methylgutaric acid (3MG) and 3hydroxyisovaleric acid (3HIVA) in body fluids. Four other inborn errors of leucine metabolism are known and in vitro 1H-NMR spectra of body fluids have been reported for several of these [3;24;27;32;39], but not for 3MGH deficiency. In vivo MR spectroscopy has also been used for studying inborn errors of metabolism. In 1H-MR spectra of the brain, various metabolites such as creatine, glutamine, glutamate, choline, myo-inositol, N-acetylaspartylglutamate and N- 132 3-Methylglutaconic aciduria type I acetylaspartate can be observed in healthy persons. Several metabolic abnormalities have been documented with in vivo MR spectra of the brain, e.g. in patients with creatine biosynthesis defects [145], succinate dehydrogenase deficiency [213], Salla disease[214], Canavan disease [215], Sjögren Larsson syndrome [216], N-acetylaspartic acid deficiency [146] and ribose 5-phosphate isomerase deficiency [116]. In this paper, we describe in vivo and in vitro NMR spectroscopic investigations of a 61-year-old woman with a 3MGH deficiency and a leukoencephalopathy which has been progressing very slowly over more than 30 years. Furthermore, we demonstrate for the first time the presence of increased 3HIVA in the brain in this disease. Our data show that 3HIVA in human body fluids may derive from brain metabolism. Materials and Methods IN VIVO MR SPECTROSCOPY MR measurements were performed on a 3.0 T MR spectrometer (Magnetom Trio; Siemens Medical Solutions, Erlangen, Germany), using the standard head coil. T2weighted images were acquired in three perpendicular planes using a TGSE sequence with an echo time of 110 ms a repetition time of 2500 ms, turbofactor 11 and an EPI factor of 3. Two-dimensional proton MRSI was carried out using PRESS volume pre-selection of a transversal slice through the ventricles with WET water suppression. Sequence parameters were TR= 2500 ms, TE= 30 ms, slice thickness= 15 mm, field-ofview= 220mm and spectral width= 1 kHz. The Kspace was sampled in a weighted, elliptical way (20 × 20 points) followed by zerofilling to 32 × 32 points and multiplication using a 50% Hamming filter before Fourier transformation. This results in an apparent voxel size of 0.7 cm3 and an effective voxal size of 2.7 cm3. Time domain data (1024 points) in the spectroscopic domain were processed using the standard Siemens software, which included removal of the residual water signal, multiplication by a Hanning filter (center= 0 ms, width= 400 ms), zero-filling to 2K data points, Fourier transformation, baseline correction, phase correction and curve fitting using Gaussian lineshapes. IN VITRO NMR SPECTROSCOPY Authentic standards. All authentic standards of the metabolites accumulating in the investigated inborn errors of metabolism, except isovalerylglycine, 3-methylcrotonylglycine, 3MGA and 3MA, were commercially available and were purchased at Sigma. 133 Chapter 3.3 Sample preparation. Plasma and cerebrospinal fluid (CSF) samples were deproteinised using a 10 kDa filter (Sartorius). Before use, the filter was washed twice by centrifugation of water to remove glycerol. Urine was centrifuged before analysis. An aliquot (20 µl for plasma and CSF; 70 µl for urine) of 20.2mM trimethylsilyl-2,2,3,3tetradeuteropropionic acid (TSP, sodium salt; Aldrich) in D2O was added to 700 µl of the ultrafiltrate or urine, providing a chemical shift reference (δH= 0.00 ppm), a concentration reference and a deuterium lock signal. The pH of each sample was adjusted to 2.50 ± 0.05 with concentrated HCl. Finally, 650 µl of the sample were placed in a 5mm NMR tube (Wilmad Royal Imperial). One-dimensional 1H-NMR spectroscopy. Body fluid samples were measured at 500MHz on a Bruker DRX 500 spectrometer with a triple-resonance inverse (TXI) 1H {15N, 13C} probehead equipped with X,Y,Z gradient coils. 1H spectra were acquired as 128 transients in 32K data points with a spectral width of 6002 Hz. The sample temperature was 298K; the H2O resonance was presaturated by single-frequency irradiation during a relaxation delay of 10 s; a pulse width of 7 µs was used (corresponded to a 90º excitation pulse). Shimming of the sample was performed automatically on the deuterium signal. To improve the spectral resolution, the samples were spun (7 Hz) during the measurements. The resonance linewidths for TSP and metabolites were <1 Hz. A π/2-shifted sine-bell window function was applied to the FID. Fourier transformation was performed after zero-filling to 64K data points. The phase and the baseline were corrected manually. The 3HIVA, cis 3MGA, trans 3MGA resonances and the TSP singlet (nine equivalent protons) were fitted semiautomatically with Lorentzian line shapes and concentrations were calculated from the relative integrals of the fitted lineshapes using the known concentration of TSP. 1H–13C HSQC spectroscopy. An heteronuclear singlequantum correlation (HSQC) spectrum of the patient’s urine was measured at 500MHz on a Bruker AV 500 spectrometer using a tripleresonance inverse CryoProbe TM (Bruker), wherein the preamplifier and the RF coils are cooled to low temperature. The sequence was optimized for 1JCH=130 Hz. FIDs (1536 transients) were acquired with 2K time-domain data points (t2), an acquisition time of 0.128 s, TR=2.128 s and 96 t1 increments. The spectral widths were 16 ppm (1H) and 70 ppm (13C), and the total measurement time was 5 days and 10 h. 1H–1H correlation (COSY) spectroscopy. A magnitude-mode COSY spectrum of the patient’s urine was acquired on a 500MHz Bruker Avance spectrometer equipped with a TCI cryo probe. The spectral 134 3-Methylglutaconic aciduria type I widths in the F1 and F2 domains were 5500 Hz. A total of 2K data points were collected in t2, 256 t1 increments with 32 transients per increment were used, and the relaxation delay was set to 2 s. Before the Fourier transformation, a sine-bell function was applied in both time domains. During the relaxation delay, the water resonance was presaturated. CASE REPORT The clinical history of the patient and the findings at MR imaging of the brain will be reported in more detail separately. The patient is a 61-year-old woman with leukoencephalopathy progressing slowly over more than 30 years, manifesting as optic neuropathy, mild ataxia and mild spasticity of the legs. For a long time the patient was misdiagnosed as suffering from multiple sclerosis. The patient did not have oligoclonal immunoglobulin bands in the CSF. T2-weighted MR images showed extensive white matter lesions of the patient’s brain. THE DIAGNOSIS The diagnosis of 3-methylglutaconic aciduria type I (OMIM: 250950) was confirmed at the metabolite level (gas chromatography for organic acid analysis of urine; NMR spectroscopy of body fluids), at the enzyme level (3MGH deficiency in fibroblasts) and the molecular genetic level. The 3MGH activity in homogenates of the patient’s fibroblasts was undetectable (controls; 2.1±0.7 nmol/min/mg protein). Analyses of the AUH gene of the patient revealed a heterozygous 559G→A (G187S) and a heterozygous 650G→A (G217D) missense mutation in exons 5 and 6, respectively. Other forms of 3-methylglutaconic aciduria, namely types II–IV, were excluded on clinical and biochemical grounds. Lactic acid in body fluids of the patient was unremarkable. Results IN VIVO NMR SPECTROSCOPY MR imaging of the brain revealed a leukoencephalopathy. Therefore, in vivo MR spectroscopic imaging was performed of the patient’s brain. Four MRSI voxels positioned in the posterior lateral subcortical periventricular white matter are shown in Figure 2. Resonances from total creatine (creatine and phosphocreatinine; N-CH3 at 3.04, N-CH2 at 3.95 ppm), GABA, glutamine and glutamate (2.19–2.4 ppm), total cholines [N-(CH3)3 group of choline, phosphocholine and glycerophosphocholine at 3.2 ppm], myo-inositol (H1,3 at 3.53 ppm, H4,6 at 3.61 ppm) and N-acetylaspartate (NAc at 2.02 ppm, Hβ1 at 2.50 ppm, Hβ2 at 2.68 ppm) were observed. Compared with voxels 1 and 2, voxel 3 showed a reduced NAA:tCho signal ratio, as often seen in patients with leukoencephalopathy [217;218]. In all spectra a singlet resonance at 1.28 ppm with an unusually small line-width of 7Hz was observed. Acyl chain protons in 135 Chapter 3.3 mobile lipids and methyl protons of lactic acid are known to contribute to resonances in this area. It is unlikely that lactic acid causes the peak at 1.28 ppm because the methyl protons of lactic acid resonate as a doublet at 1.33 ppm. Lipid signals are known to resonate at 1.3 ppm (methylene protons) and 0.85 ppm (methyl protons). However, generally, the lipid methylene signals have a broader line width than the 7 Hz observed in the patient’s brain. Figure 2 A 3.0T T2 weighted brain image (upper right) and four MRSI voxels (apparent voxel size=0.7 cm3) positioned in the posterior lateral subcortical periventricular white matter (upper left) showing a resonance of 3-hydroxyisovaleric acid. Compared with voxels 1 and 2, voxel 3 (lower) showed a reduced NAA:tCho signal ratio. Furthermore, the spectrum contains resonances of the normal brain metabolites total creatines (tCr), myo-inositol (myo-Ins), γ-aminobutyric acid (GABA), glutamine and glutamate (Glx) In addition, the resonance at 0.85 ppm visible in Figure 2 may also originate from methyl groups of proteins and amino acids. Furthermore, no lipid contamination originating from the skull region is expected to contribute to the spectra from this location. Still, we cannot rule out the possibility that some lipid signal contributes to the intensity at 1.28 ppm. It was then discovered by in vitro NMR spectroscopy that the patient had an abnormally high concentration of 3HIVA in her CSF (172 µmol/l, reference value, <10µmol/l). At pH 2.5, 3HIVA gives a singlet resonance at 1.33 ppm (six protons of the two methyl groups) and a singlet resonance at 2.55 ppm (two protons of the methylene group). Since brain pH values range from 6.96 to 7.05 [219], a one-dimensional (1D) 1H-NMR spectrum of a reference solution of 3HIVA in H2O was recorded at pH 7.0. The singlet resonance from the six 136 3-Methylglutaconic aciduria type I equivalent protons of the two methyl groups shifted from 1.33 ppm at pH 2.5 to 1.27 ppm at pH 7.0. These data provide evidence supporting the assignment of 3HIVA as the source of the resonance at 1.28 ppm in the in vivo MR brain spectrum. Owing to overlap of glutamate/glutamine (Glx) the two equivalent protons of the methylene group of 3HIVA (singlet: 2.37 ppm at pH 7) could not be observed in the in vivo MR brain spectrum. The singlet resonance at 1.28 ppm and the total creatine (tCr) singlet resonance (3.04 ppm; three equivalent protons) in the in vivo spectrum were fitted semiautomatically with Gaussian line shapes. Assuming that only 3HIVA and no lipid contribute to the peak at 1.28 ppm then the concentration of 3HIVA can be estimated from the relative integrals of the fitted lineshapes. Based on the reference value of 5.5 mmol/l for total creatine in occipital white matter [150], 3HIVA was calculated to be 0.9 mmol/l, about five times the concentration in CSF. IN VITRO NMR SPECTROSCOPY Expanded regions (1H-axis: 1.00–3.25 ppm/13C-axis: 15.0–55.0 ppm) of the 1D 1H and the two-dimensional (2D) 1H–13C HSQC 2D NMR spectra of the patient’s urine at pH 2.5 are shown in Figure 3. The major metabolites in this region and the 1H/13C shift values for the correlation peaks are creatinine (singlet: 3.13/38.8), creatine (singlet 3.05/40.0), citric acid ([AB]2 spin system 2.99; 2.83/46.5), dimethyl amine (triplet 2.71/37.8), acetic acid (singlet 2.08/23.5), alanine (doublet 1.51/18.7) and lactic acid (doublet 1.41/22.4). Furthermore, unusually high concentrations of 3HIVA, cis 3MGA, trans 3MGA and 3MG were observed in the urine spectra. These characteristic 3-methylglutaconic aciduria metabolites were assigned and quantified as described below, and their structures and NMR data are summarized in Table 1. Figure 3 500 MHz 1D 1H spectrum (upper panel) and 2D 1H-13C HSQC spectrum (lower panel) of the patient’s urine (pH=2.5). The insert shows the region between 2.2 and 1.9 ppm in greater detail. 137 Chapter 3.3 3HIVA 1H-NMR of model compounds in H2O at pH 2.5 revealed that the two unusually intense singlets at 1.33 and 2.55 ppm, observed with integral ratio 3:1 for the patient’s urine (Figure 3), derived from 3HIVA (Table 1). In two different urine samples of the patient the concentration of 3HIVA amounted to 63 and 61µmol/ mmol creatinine [control urines from normal adults (n=12), mean 8, range 3–15]. Also CSF and plasma of the patient showed increased concentrations of 3HIVA (CSF, 172 µmol/l, reference value <10 µmol/l; plasma 20 µmol/l, reference value <10 µmol/l). Table 1: NMR assignments (urine, pH 2.5) for metabolites characteristic of 3methylglutaconic aciduria, type I Pos. δH a Mult. δC 2 4, 4’ 2.55 s 50.2 3-Methylglutaric acidC 3 6 2.35 m 3-Methylglutaconic 2 4 6 3.65 d 41.9 41.4 5.96 m 121.9 115.4 2 4 6 3.28 d 48.4 47.4 5.85 m 122.2 115.4 Metabolite 3-Hydroxyisovaleric acid acid “Cis” 3-Methylglutaconic acid “Trans” Structure 1.33 1.01 1.99 2.14 s δ Cb 31.1 d d d 28.6 21.6 22.2 14.2 a Chemical shifts in ppm rel. TSP; multiplicities are: s, singlet; d, doublet; m, multiplet. shift predictions made by ChemOffice 2004. c Spectrum not yet completely interpreted. b Chemical Cis- and trans-3MGA 3MGA is not commercially available as a model compound. Therefore, 1D 1H, 2D COSY and 2D HSQC experiments on a urine sample of the patient were used to make the assignments for the spin systems of the two isomers of this metabolite (Table 1). It is expected that the carboxyl group attached to the C4 of the double 138 3-Methylglutaconic aciduria type I bond will exert a downfield shift on methyl or methylene (C6 or C2) in a cis position compared with the trans position relative to the carboxyl. Therefore, the doublet resonances at 1.99 (4J4,6=1 Hz) and 2.14 (4J4,6= 1 Hz) ppm were assigned to the methyl protons of the cis and trans forms of 3MGA, respectively. With this assignment of the isomers the methylene protons on C2 are downfield-shifted for the cis isomer as expected, and the order of shifts for H4, C2 and C6 agree with predictions made by ChemOffice 2004 software. For example, the predicted and observed 13C shift for C2 are 41.4 and 41.9 ppm (cis-3MGA) and 47.4 and 48.4 ppm (trans-3MGA). These assignments for the 3MGA isomers agree with those of Iles et al. [32] Lehnert and Hunkler [24] incorrectly assigned the H4 multiplet at 5.96 ppm to trans 3MGA and the multiplet at 5.85 to cis 3MGA. In two different urine samples of the patient the concentration of 3MGA was 94 (cis/trans values: 62/32) and 141 (cis/trans values: 90/51) µmol/mmol creatinine [normal adult range: 1.0–6.5 µmol/mmol creatinine [220]]. In CSF, only the cis form of 3MGA was observed (125 µmol/l, reference <10µmol/l). In plasma no 3MGA was observed. All samples have been adjusted to pH 2.5 with concentrated HCl so that cis/trans interconversion could be an artifact due to the pH changes. To test this a urine sample was recorded at pH 9 and pH 2.5. Both spectra showed the same cis/trans ratio of 3MGA. Thus, pH changes are unlikely to be a source of 3MGA cis/trans interconversion. 3MG 3MG is also not commercially available as a model compound. Again, 1D and 2D COSY NMR experiments on a urine sample of the patient allowed the assignments of this metabolite (Table 1). The doublet resonance at 1.01 ppm, observed in the methyl region of the 1D 1H-NMR urine spectrum, was assigned to the methyl protons of 3MG. In the COSY spectrum, this doublet showed a cross peak at 2.35 ppm (Figure 4). This multiplet resonance derives from the single CH proton of 3MG. The chemical shift of the resonances of 3MG agree exactly with those of Lehnert and Hunkler [24]. However, these authors incorrectly assigned the methyl protons of 3MG as a multiplet. Because of the low concentration of 3MG in urine, 13C assignments of 3MG in the 1H-13C HSQC spectrum were not possible. In two different urine samples of the patient the concentrations of 3MG were 15 and 12 µmol/mmol creatinine [reference adult 0.5±0.2 µmol/mmol creatinine [220]]. In CSF and plasma no 3MG was observed. 139 Chapter 3.3 Figure 4 500MHz 1H 2D COSY spectrum of the patient’s urine (pH=2.5). The cross peak of 3methylglutaric acid is shown in the insert for the selected region (gray box). OTHER INBORN ERRORS OF LEUCINE METABOLISM We measured urine samples from seven patients suffering from four other diseases involving leucine metabolism. Resonance assignments were based on previouslyreported spectra [24] and on 1D and 2D spectra of 2-hydroxyisovaleric acid, 2-oxoisovaleric acid, 3-methyl-2-oxo-valeric acid, 3-hydroxyisovaleric acid and 3-hydroxy-3-methylglutaric acid dissolved in H2O (pH 2.5). The 1D 1H-NMR spectrum of urine from a patient with maple syrup urine disease (Figure 1, enzyme B) showed the oxo acids 3-methyl-2-oxo-valeric acid (1713 µmol/mmol creatinine), 2-oxoisovaleric acid (558 µmol/mmol creatinine) and 2oxoisocaproic acid (2951 µmol/mmol creatinine) as well as the presence of 2hydroxyisovaleric acid (2622 µmol/mmol creatinine). High concentrations of isovalerylglycine (4895 µmol/mmol creatinine) and 3HIVA (1505 µmol/mmol creatinine) could be observed in the 1H-NMR spectrum of urine from a patient with isovaleric acidemia (Figure 1, enzyme C). In two unrelated patients with 3methylcrotonyl CoA carboxylase deficiency (Figure 1, enzyme D), unusually high concentrations of 3HIVA (6400 and 6300 µmol/mmol creatinine) and 3methylcrotonylglycine (345 and 800 µmol/mmol creatinine) were observed in the urine 1D 1H-NMR spectra. Furthermore, we measured urine samples from three patients with 3-hydroxy-3-methylglutaryl-CoA lyase deficiency (Figure 1, enzyme F). In all 1H-NMR spectra we observed high concentrations of 3-hydroxy-3- 140 3-Methylglutaconic aciduria type I methylglutaric acid (3365, 2037 and 4259 µmol/mmol creatinine), 3HIVA (877, 918 and 1136 µmol/mmol creatinine), 3MG (382, 219 and 441µmol/mmol creatinine) and 3MGA (2921, 2249 and 3638 µmol/mmol creatinine). In all three urine samples a cis/trans 3MGA ratio of 2:1 was observed. Table 2 provides an overview of all five known inborn errors of leucine metabolism and their diagnostically important metabolites with the 1H chemical shift data observed in urine samples from affected patients (pH 2.5). Table 2: 1H chemical shifts (pH 2.5) for metabolites accumulating in urine for inborn errors of leucine metabolism (Figure 1; B–F) Metabolite δH and Mult. a Affected enzyme b in Figure 1 2-Hydroxyisovaleric acid 0.88d; 0.98d; 2.06m; 4.09d B 2-Oxoisocaproic acid 0.94d; 2.09m; 2.61d B 3-Methyl-2-oxovaleric acid 0.95t; 1.09d; 1.45m; 2.92m B 2-Oxoisovaleric acid 1.12d; 3.02m B Isovalerylglycine 0.94d; 2.18m; 3.94d C 3-Hydroxyisovaleric acid 1.33s; 2.55s 3-Methylcrotonylglycine 1.86d; 2.03d; 3.97d; 5.78m 3-Methylglutaconic acid (cis) 1.99d; 3.65d; 5.96m E, F 3-Methylglutaconic acid (trans) 2.14d; 3.28d; 5.85m E, F 3-Methylglutaric acid 1.01d; 2.35m E, F 3-Hydroxy-3-methylglutaric acid 1.41s; 2.71d F C, D, E, F D a Chemical shifts in ppm rel. TSP; multiplicities are: s, singlet; d, doublet; m, multiplet s, singlet; d, doublet; m, multiplet. b Disorders: B, maple syrup urine disease; C, isovaleric acidemia; D, 3-methylcrotonyl CoA carboxylase deficiency; E, 3-methylglutaconyl CoA hydratase deficiency; F, 3-hydroxy-3-methylglutaryl-CoA lyase deficiency. 141 Chapter 3.3 Discussion This study describes in vivo and in vitro NMR investigations of a 61-year-old patient with 3MGH deficiency and a slowly progressing leukoencephalopathy. In the literature 12 children between 0 and 14 years of age have been reported with 3MGH deficiency. Some papers have reported leukoencephalopathy in these patients [221]. In none of these studies was MR spectroscopy performed. In vitro NMR and in vivo MR spectra showed that our patient had high levels of 3HIVA in her CSF and probably in her brain. Several NMR experiments confirmed the presence of 3HIVA, 3MGA and 3MG in body fluids of our patient. NMR spectroscopy of body fluids can provide an overview of proton-containing metabolites and the technique can be used for quantification. Lehnert and Hunkler [24] were the first to perform NMR spectroscopy on urine samples and model compounds under standardarized pH conditions (pH 2.50). They used 1H-NMR at 250 MHz to investigate urine samples of 13 different inborn errors of metabolism, including three diseases from leucine metabolism: isovaleric acidemia, 3methylcrotonyl-CoA carboxylase deficiency and 3-hydroxy-3-methylglutaryl-CoA lyase deficiency. Iles et al. [32] reported 1D and 2D COSY 1H-NMR spectra of urine from twins with 3-hydroxy-3-methylglutaryl-CoA lyase deficiency. Until now, body fluids of patients with type I 3MGH deficiency have not been studied with NMR spectroscopy. The combination of 1D 1H, 2D 1H-1H COSY and 1H-13C HSQC spectra allowed identification of the resonances of 3MGA, 3MG and 3HIVA in urine. Moreover, 1D spectra were used to quantify these metabolites. This led to the correct diagnose of 3-methylglutaconic aciduria. 3-Hydroxy-3-methylglutaric acid was not detectable in urine, thus excluding 3-hydroxy-3-methylglutaryl- CoA lyase deficiency. We had the opportunity to study urine samples of seven patients with four other inborn errors of leucine metabolism. As shown in Table 2, for all diseases NMR spectroscopy allowed identification of the diagnostically important metabolites and the correct diagnosis. In most studies, only data obtained from urine samples have been reported. We had the opportunity to analyze plasma as well as CSF from our patient. In plasma, slightly increased 3HIVA was the single unusual finding observed in theNMR spectrum. This may be explained by the rapid and efficient renal clearance of 3HIVA, 3MGA and 3MG. In an earlier paper we reported a 3HIVA concentration of 540 µmol/l in CSF from a patient with 3-methylcrotonyl-CoA carboxylase deficiency [3] (Figure 1, enzyme D). This concentration was about three times higher than in the CSF of our current patient (172 µmol/l). The CSF:plasma ratio of 3HIVA for our patient was about 8.6. Duran et al. reported a CSF:plasma 3HIVA ratio of ca. 1.7 in one patient with 3-methylcrotonyl-CoA carboxylase deficiency [222] (Figure 1; enzyme D). These data may indicate that CSF 3HIVA in our patient derives from brain leucine metabolism (Figure 1). Moreover, in vivo MR spectra suggested that our patient had high levels of 3HIVA in her brain. Yudkoff et al. [223] have demonstrated the transamination of leucine 142 3-Methylglutaconic aciduria type I into 2-oxoisocaproic acid in rat brain and Bixel et al. [224;225] showed the presence of branched-chain aminotransferase (Figure 1; enzyme A), branchedchain a-keto acid dehydrogenase (Figure 1; enzyme B) and b-methylcrotonyl-CoA carboxylase (Figure 1; enzyme D) in astroglial cells from rats [225]. This suggests that leucine, which is involved in the synthesis of the neurotransmitter glutamate [223], can be degraded into acetoacetate in the human brain (Figure 1). However, to our knowledge, the presence of 3MGH activity in brain cells has not been described in the literature. Biotinidase deficiency (OMIM: 2532620) is a form of multiple carboxylase deficiency in which the formation of the cofactor biotin from biocytin via biotinidase is inhibited. Patients with this disease may have high concentrations of 3HIVA, 3-methylcrotonylglycine, methylcitric acid and lactic acid in their body fluids. Biotinidase deficiency usually presents with neurological problems. Three symptomatic patients had CSF/plasma 3HIVA ratios ranging between 4.3 and 8.4 [222]. In a fourth patient the CSF/plasma ratio was 27.8. This patient had severe chronic neurological problems. The authors speculated that 3HIVA may have a specifically harmful effect on the brain. Our study shows an accumulation of 3HIVA in the brain of our patient with 3-methylglutaconic aciduria while we found no evidence for accumulation of 3MGA in the in vivo spectrum. However, we cannot exclude an overlap of 3MGA resonances with other resonances in the in vivo spectrum. The concentration in CSF of 3HIVA was higher than the concentration of 3MGA in our patient. The leukoencephalopathy in our patient is in line with the hypothesis that 3HIVA may be the key factor in the pathogenesis of central nervous system pathology in biotinidase deficiency and 3-methylglutaconic aciduria. 1D 1H NMR spectroscopy can be used to quantify both the cis and trans form of 3MGA in body fluids. 3-Methylcrotonyl CoA carboxylase (Figure 1; enzyme D) is a stereospecific enzyme and produces only trans 3-methylglutaconyl-CoA. Lynen et al. [226] reported isomerization of cis/trans 3-methylglutaconyl-CoA under alkaline conditions and Kelley et al. [220] showed that the choice of conditions for derivatization by GC-MS might influence the cis/trans ratio of the free acid. A cis/trans 3MGA ratio of 2:1 was observed in all urine samples from the three patients with 3-hydroxy-3-methylglutaryl-CoA lyase deficiency. Iles et al. reported a similar ratio in urine NMR spectra of twins with 3-hydroxy-3methylglutaryl-CoA lyase (Figure 1; enzyme F) deficiency [32], which we could confirm in this study on three patients with this defect. In our patient with 3MGCoA hydratase deficiency, a cis/trans 3MGA ratio of ca. 2:1 was observed in both urine samples, while in CSF only cis-3MGA was observed. Different cis/trans ratios for 3MGA could be the result of spontaneous cis/trans interconversion of the free acid according to the environment in different organs or body fluids. However, this patient’s urine NMR spectra recorded at pH 9 and pH 2.5 showed the same cis/trans 3MGA ratio. This suggests that the pH is unlikely to be a source of 143 Chapter 3.3 cis/trans interconversion. Another explanation for the occurrence of only cis 3MGA in CSF may be the presence of a brain-specific isoform of 3-methylcrotonyl CoA carboxylase which may have stereospecificity. As yet, there is no other evidence for a separate brain isoform. Further investigations are necessary to explain this finding. In conclusion, we have shown that NMR spectroscopy can be used to identify and quantify the characteristic 3-methylglutaconic aciduria metabolites 3MGA, 3MG and 3HIVA in body fluids. Furthermore, our data suggest that 3HIVA accumulates in brain of a patient with 3-methylglutaconic aciduria, type I. Our data show that 3HIVA and 3MGA may derive from brain metabolism. Acknowledgements The authors thank Dr Jörn Oliver Sass (Stoffwechsellabor, Zentrum für Kinderheilkunde- und Jugendmedizin, Universitätsklinikum Freiburg, Germany) for providing urine samples from patients with 3-hydroxy-3-methylglutaryl-CoA lyase deficiency. 144 CHAPTER 3.4 Leukoencephalopathy in adult-onset 3-methylglutaconic aciduria type I Berry Kremer, Udo F.H. Engelke, Janneke Schuuring, Marinette van der Graaf, Ference J. Loupatty, Ronald J.A. Wanders, Jouke Mellema and Ron A. Wevers Submitted for publication 145 146 Leukoencephalopathy in adult-onset 3-methylglutaconic aciduria type I Abstract We describe a 61-year-old woman with 3-methylglutaconic aciduria type I (3methylglutaconyl-CoA hydratase deficiency) and a very slowly progressive leukoencephalopathy over more than 30 years, manifesting as optic neuropathy, mild ataxia and leg spasticity. Organic acid analysis and in vitro NMR spectroscopy of body fluids revealed elevated 3-methylglutaconic acid, 3-methylglutaric acid and 3-hydroxyisovaleric acid suggesting 3-methylglutaconyl-CoA hydratase deficiency as confirmed enzymatically. In vivo MR spectroscopy revealed an increase of brain intraparenchymal 3-hydroxyisovaleric acid. Mutation analysis demonstrated heterozygosity for two novel missense mutations (559G→A; 650G→A) in the AUH gene (3-methylglutaconyl-CoA hydratase). This patient is the oldest patient ever described with 3-methylglutaconic aciduria type I. 147 Chapter 3.4 Introduction 3-Methylglutaconic aciduria comprises a group of metabolic disorders biochemically characterized by elevated urine levels of 3-methylglutaconic acid (3MGA) and 3-methylglutaric acid (3MG). To date, four distinct forms have been recognized. 3-Methylglutaconic aciduria type I (OMIM 250950) is caused by a deficiency of 3methylglutaconyl-CoA hydratase (3MGH, E.C. 4.2.1.18) [227]. This enzyme catalyzes the fifth step in leucine catabolism, which is the conversion of 3methylglutaconyl-CoA to 3-hydroxy-3-methylglutaryl-CoA. The hydratase deficiency is associated with symptoms varying from mildly delayed speech development to neurologic, cardiac and hepatic abnormalities that until now have been described in young children, only. Patients with other forms of 3methylglutaconic aciduria (Barth syndrome, OMIM 302060; Costeff syndrome, OMIM 258501; unspecified 3-methylglutaconic aciduria, OMIM 250951) have normal hydratase activity [227]. 3MGH is encoded by the AUH gene [228;229]. Mutation analysis of AUH in seven reported patients revealed protein function abolishing mutations [221;227-229]. We describe a patient with a protracted adult onset 3-methylglutaconic aciduria type I, heterozygous for two novel missense mutations in AUH. As far as we are aware, this is the first adult patient ever diagnosed with this condition. Case Report Since age 35, after giving birth to her second child, this woman had experienced a progressive visual decline, initially with recurrent episodes of rapid deterioration and subsequent improvement. Initial neurological and ophthalmologic examinations, at age 37, revealed a visual acuity of 0.6 and 0.8 of the right and left eye, respectively; optic disc pallor and atrophy bilaterally, slight gait and limb ataxia and bilateral extensor responses of the toes. The electroretinogram indicated rod dysfunction. CSF studies revealed no abnormalities. Over the years, her visual acuity gradually deteriorated with fluctuations. Repeated Goldmann perimetry showed progressive concentric visual field constriction and an expanding blind spot. Ataxia and pyramidal abnormalities increased slightly, while orobucal dyskinesia was noted at the age of 40 years. A neurological re-evaluation at age 61, revealed: preserved light / dark perception; severe optic disc pallor; slight pendular nystagmus; minimal dysarthria; mild limb ataxia with severe gait ataxia and frequent falls despite preserved ambulation; increased tendon reflexes with absent ankle tendon reflexes and extensor responses of the toes. Nerve conduction studies at that age were still normal but routine 1.0 Tesla MR-imaging revealed marked supratentorial symmetric confluent subcortical white matter abnormalities (Figure 1A demonstrates a subsequent 3.0 Tesla image). Routine 148 Leukoencephalopathy in adult-onset 3-methylglutaconic aciduria type I biochemical and hematological analyses were completely normal. The CSF contained a normal cell count, a total protein content of 392 mg/l, lactate of 1731 µmol/l, without intrathecal immunoglobulin synthesis and no abnormal IgG bands. Figure 1 In vivo MR spectroscopic imaging. T2-weighted MR images show extensive white matter lesions of the patient’s brain (Panel A). The in vivo spectrum of an MRSI voxel (3T, PRESS, TE/TR= 30/2500 ms) positioned in the posterior lateral subcortical periventricular white matter is shown in Panel B. (The voxel position is marked in white rectangles (Panel A)). The spectrum shows an unusual singlet resonance deriving from 3-hydroxyisovaleric acid. Furthermore, the spectrum shows a reduced NAA/tCho signal ratio as often seen in patients with leukoencephalopathy. Analysis of organic acids in the urine of the patient repeatedly showed strongly elevated 3MGA and 3-hydroxyisovaleric acid (3HIVA) and increased 3MG. Using proton NMR spectroscopy with a 500 MHz Bruker spectrometer both the cis and the trans form of 3MGA could be consistently demonstrated in urine (in two samples the cis-form: 62 and 90 µmol/mmol creatinine and the trans-form: 32 149 Chapter 3.4 and 51 µmol/mmol creatinine; reference for total 3MGA 1.0-6.5 for adults [220]). Details on the identification of cis and trans 3MGA with body fluid NMR spectroscopy have been published elsewhere [47]. Urinary 3MG amounted to 15 and 12 µmol/mmol creatinine in two samples (reference 0.5±0.2 µmol/mmol creatinine [220]). In addition 3HIVA was increased (63 and 61 µmol/mmol creatinine; reference < 15 µmol/mmol creatinine). The in vitro NMR spectrum of the CSF revealed the presence of only the cis-form of 3MGA (125 µmol/l; not detectable in normal CSF). This suggests that the transform is not synthesized in the brain or is immediately converted to the cis-form. 3HIVA in CSF amounted to 172 µmol/l (reference <10). 3MGA could not be detected in plasma of the patient and the plasma concentration of 3HIVA (20 µmol/l; reference <10) was substantially lower than in CSF, suggesting that both compounds are synthesized in the brain of the patient. Figure 2 compares the proton NMR spectra of urine and CSF of the patient. In vivo brain MR spectroscopy on a 3.0 Tesla MR imaging system revealed a resonance at 1.28 ppm in our patient with an unusual small line width of 7 Hz. This resonance was present in abnormal white matter and is absent in normal brains (Figure 1B). At physiological pH 3HIVA resonates at 1.28 ppm [47]. As lipid resonances usually have a broader line width the unusual resonance at 1.28 ppm may be caused by 3HIVA. Assuming that no lipids contribute to the signal at 1.28 ppm the estimated concentration of 3HIVA in the brain of our patient is 0.9 mmol/l which is about five times higher than in the CSF. To determine whether the elevated urinary metabolite levels were caused by a deficiency of 3MGH we measured the activity of this enzyme in cultured human skin fibroblasts and in lymphocytes as described elsewhere using 20-50 mg/l of lymphocyte protein. 3MGH activity [230] in fibroblasts and lymphocytes of the patient was undetectable (reference values fibroblasts 2.1 ± 0.7 nmol/min/mg protein and lymphocytes 1.8 ± 0.2 nmol/min/mg protein). In addition, we performed mutation analysis for AUH on genomic DNA derived from fibroblasts and could identify a heterozygous 559 G>A (G187S) and a heterozygous 650 G>A (G217D) missense mutation in exons 5 and 6, respectively. 150 Leukoencephalopathy in adult-onset 3-methylglutaconic aciduria type I Figure 2 Body fluid 1H-NMR spectra of the patient’s CSF and urine at 500 MHz (A and B, respectively). The inserts show the region between 2.2 and 1.95 ppm in greater detail. 151 Chapter 3.4 Discussion This case demonstrates four features of 3-methylglutaconic aciduria type I. The patient is unique in terms of age of presentation and perhaps of onset of her disease; 3-methylglutaconic aciduria type I may present as a leukoencephalopathy; in vivo MR spectroscopy may demonstrate increased brain tissue levels of 3-hydroxisovaleric acid; and we describe two novel mutations. 3Methylglutaconic aciduria type I (3MGH deficiency) is a very rare recessive disorder of leucine metabolism that usually presents in infancy or early childhood with either severe metabolic derangements, or, alternatively, as a mild psychomotor retardation. Until now 12 patients have been reported, with 1 case diagnosed at neonatal screening and asymptomatic at age 2½ years [221;228;231-239]. From the literature the natural course of the disease is unclear. Our patient suggests that if affected individuals survive acute metabolic crises in infancy, the disease may develop as a very slowly progressive neurodegenerative disorder with perhaps a normal or only slightly reduced life expectancy. Its ultimate phenotype may be a progressive demyelinating leukoencephalopathy that in our patient had long been misdiagnosed as multiple sclerosis. Although MRI reports describe predominant basal ganglia gray matter lesions [231], white matter lesions have been mentioned as well [221;231;232;238] The leukoencephalopathy in our patient may be caused by 3MGA. Alternatively 3HIVA may be the causative factor. Duran has speculated about a toxic effect of accumulating 3HIVA on the brain in patients with biotinidase deficiency and patients with 3-methylcrotonyl CoA carboxylase deficiency [222]. The obvious accumulation of 3HIVA in the brain of our patient is a further clue towards the potential toxic effects of this metabolite on the brain. An intriguing finding in our patient is the severe optic neuropathy, probably due to demyelination as well. Optic neuropathy has been described as a feature of 3methylglutaconic aciduria type III (Costeff syndrome). The urinary 3MGA concentration is lower in Costeff patients than in type I cases. Costeff syndrome is caused by a mutation in the OPA3 gene [240]. 3HIVA does not accumulate in Costeff syndrome. The function of the protein encoded by the OPA3 gene is unknown. Therefore the optic neuropathy in Costeff syndrome may either be due to the accumulating 3MGA and/or the 3MG or to an as yet unknown metabolic consequence of the defect in the OPA3 gene. We demonstrate that MR spectroscopy may reveal an increase in 3HIVA in 3MGH deficiency. Resonances from 3MGA and 3MG were not observed in the in vivo spectrum but these may have been obscured by other, predominant, metabolites. AUH encodes 3MGH that catalyzes the conversion 3-methylglutaconyl-CoA to 3hydroxy-3-methylglutaryl-CoA [227]. Until now, 6 mutations have been described: one exon 9 skipping splice acceptor mutation; a mutation in the second nucleotide of the intron 9 splice donor sequence; a missense mutation, a point deletion and an insertion that all three introduce premature stop codons and an 152 Leukoencephalopathy in adult-onset 3-methylglutaconic aciduria type I exon 7 missense mutation [221;228;229]. Mutation analysis of our patient’s genomic DNA revealed two novel missense mutations, substituting two glycine residues at position 187 and 217 by the amino acids serine and aspartate, respectively. BLAST analysis with the mRNAs of human AUH [Genbank: NM_001698], human short-chain enoyl-CoA hydratase [Genbank: NM_004092] and the hydratase domain of the human mitochondrial trifunctional protein [Genbank: NM_000182] revealed at least 400 related proteins of the hydratase family of different origins. All of these proteins contained conserved glycine residues at positions 187 and 217, suggesting that the glycines at positions 187 and 217 are critical for enzyme function. Indeed, hydration of an enoyl-CoA substrate by 3MGH proceeds through an intermediate which is stabilized via hydrogen bond interactions with two peptide NH-groups [241]. These amino groups are positioned in two distinct, but conserved regions that are an intrinsic part of the active site pocket. Glycine 187 of 3MGH is located in one these regions. Mutagenesis studies of this region in human short-chain enoyl-CoA hydratase revealed that substitution of glycine with proline resulted in a 106-fold decrease in hydratase activity, whereas the overall protein production of this mutant remained unaffected [242]. One can easily imagine that substition of glycine with a polar serine residue, as described in our patient, will result in a similar effect and, thus, a strongly decreased activity of 3-methylglutaconyl-CoA hydratase. The three-dimensional structure of 3-methylglutaconyl-CoA hydratase is a dimer of trimers [243]. Glycine 217 is located in a short flexible region (7 nucleotides) which is surrounded by two α-helices; one helix contains a catalytic glutamate residues, whereas the other is involved in subunit interactions to form the trimers. Both the catalytic activity and trimer formation depend on hydrogen bond interactions. Substitution of glycine 217 with the charged amino acid residue aspartate could alter the conformation of the flexible region thereby disrupting the distance required for optimal hydrogen bond interaction. As a result, trimer formation and catalytic activity could be restrained. Indeed, in lysates of fibroblasts and lymphocytes from our patient we have found no residual 3MGH activity. This strongly suggests that these mutations abolish the activity of the enzyme. Expression studies have to be done in the future to prove this unambiguously. 153 154 FOUR LIPID METABOLISM 155 156 CHAPTER 4.1 Diagnosing inborn errors of lipid metabolism using 1H Nuclear Magnetic Resonance spectroscopy Udo F.H. Engelke, Marlies Oostendorp, Michèl A.A.P. Willemsen and Ron A. Wevers Clinical Chemistry 2006; 52: 1395 – 1405 157 158 Inborn errors of lipid metabolism Abstract Background: Many severe diseases are caused by defects in lipid metabolism. As a result, patients often accumulate unusual lipids in their blood and tissues and proper identification of these lipids is essential for correct diagnosis. In this study, we investigated the potential use of proton nuclear magnetic resonance (1H-NMR) spectroscopy to simultaneously identify and quantify (un)usual lipids present in the blood of patients suffering from different inborn errors of lipid metabolism. Methods: Blood plasma or serum lipids were extracted in 2:1 (v/v) chloroform- methanol. After addition of the non-volatile chemical shift and concentration reference compound octamethylcyclotetrasiloxane, 1H-NMR measurements were performed on a 500 MHz spectrometer. Assignments were based on literature, computer simulations and reference spectra of relevant authentic standards. Results: Spectra of normal plasma samples allowed the identification of nine lipid species. For cholesterol and triglyceride levels, a good correlation was found between conventional methods and 1H-NMR. Furthermore, four inborn errors of lipid metabolism were investigated (three in sterol metabolism and one in fatty acid metabolism). NMR analysis led to a correct diagnosis for all four diseases, while the concentration of the diagnostic metabolite could be determined for three. Conclusions: We showed that 1H-NMR spectroscopy of blood plasma or serum lipid extracts can be used to accurately identify lipids. Identification was also possible for unusual lipids in the blood of patients suffering from an inborn error of lipid metabolism. Additionally, simultaneous lipid quantification was possible for the first time. The technique is therefore applicable in clinical diagnosis and follow up. 159 Chapter 4.1 Introduction Proton nuclear magnetic resonance (1H-NMR) spectroscopy is a versatile technique that can be used in a wide range of disciplines. The best known medical applications are in vivo magnetic resonance imaging and spectroscopy, but in vitro 1H-NMR spectroscopy of body fluids has also been used to diagnose inborn errors of metabolism (e.g. [3;4;11]). In contrast to conventional techniques, NMR can detect the majority of all proton containing metabolites in a single experiment of approximately 15 minutes. Furthermore, it is a non-destructive method and requires little to no sample preparation. Currently, blood plasma or serum samples are either measured directly [7;23;244] or after deproteinization by ultrafiltration [4]. The advantage of using the latter method is that broad, overlapping protein resonances are removed, thereby yielding a highly resolved spectrum in which only the water soluble, low molecular mass metabolites are observed. Unfortunately, this also limits NMR as a diagnostic tool for inborn errors of metabolism, to diseases involving accumulation or absence of these relatively small metabolites. There are also many severe diseases that are caused by inherited defects in the metabolism or biosynthesis of different fatty acids and sterols. As a result, unusual lipids are often present in the blood and tissues of affected patients. Accurate identification and quantification of these metabolites is essential for correct diagnosis of the disease. 1H-NMR has several major advantages over GC-MS and LC-MS for analysis of lipids and sterols. First, authentic standards are usually not required once the chemical shifts of the biologically relevant species are known (see [245] for the chemical shifts of many C27-sterols and their acetyl derivatives). Secondly, lipid identification is almost unequivocal if a few distinct resonances are resolved. Finally, sample preparation for 1H-NMR measurements can be fairly simple, whereas the conventional biochemical analysis of unusual lipids in body fluids can be complicated and time-consuming, sometimes involving derivatization steps and a combination of several types of chromatography. Using one-dimensional (1D) 1H-NMR spectroscopy of intact blood plasma, several lipid signals can be detected [23]. The technique can be used to determine the relative amounts of high, low and very low density lipoprotein cholesterol using complex mathematical line fitting techniques [244]. However, the diagnostic markers for inborn errors in lipid metabolism remain undetected due to their low concentrations, overlap with other metabolites and protein derived interferences. Here, we describe a simple procedure based on the Folch-extraction [246] to isolate all lipids from blood plasma or serum samples using a 2:1 (v/v) chloroform-methanol extraction medium. Casu et al. [247] reported the NMR analysis of lipids extracted from erythrocytes and plasma of humans. Furthermore, 1H-NMR spectroscopy has been applied successfully in the diagnosis of the SmithLemli-Opitz syndrome [248;249]. This study demonstrates for the first time that 160 Inborn errors of lipid metabolism several serum lipids can be quantified simultaneously using 1H-NMR spectroscopy. The clinical usefulness of the technique is shown by successfully applying it to four inherited disorders in lipid metabolism (Smith-Lemli-Opitz syndrome, cerebrotendinous xanthomatosis, sitosterolemia and Refsum disease). Materials and Methods AUTHENTIC STANDARDS All authentic standards of the metabolites accumulating in the investigated inborn errors of metabolism, except 8DHC and 8-lathosterol, were commercially available and were purchased at Sigma. Authentic standards of cholesterol, 7-lathosterol, oleic acid, linoleic acid, palmitic acid and stearic acid were purchased at Sigma as well. SAMPLE PREPARATION Blood plasma or serum samples were obtained from healthy controls and patients suffering from an inborn error of lipid metabolism at the University Medical Centre Nijmegen, The Netherlands. All samples were provided anonymously after routine diagnostic screening was performed and were kept frozen until NMR analysis. All lipid material was extracted from 1 mL of blood plasma or serum using a Folch extraction [246] optimized for blood plasma [250]. Per extraction, 30 mL of a 2:1 (v/v) chloroform-methanol extraction medium was used. The extraction was performed in a 50 ml capped Teflon® centrifuge tube (Nalgene, Rochester, NY). The water-methanol layer and denaturated protein precipitate were discarded, while the chloroform layer was evaporated to dryness using an AS290 automatic Speedvac concentrator (Savant Instruments, Farmingdale, NY). Subsequently, the extract was redissolved in approximately 650 µl CDCl3 for NMR analysis. The use of fresh chloroform and methanol during sample preparation is important. Aged chloroform can contain phosgenes, which can react with the analytes and hence lead to incorrect results. Unfortunately, the conventional chemical shift reference compound tetramethylsilane (TMS) was not suitable as concentration reference due to its high volatility. Therefore, octamethylcyclotetrasiloxane (OMS; Fluka), which has a boiling point of 448 K and a chemical shift of 0.094 ppm compared to TMS, was used instead. Pipetting an amount of OMS dissolved in CDCl3 to the sample could not be performed, because the use of plastic pipet tips with chloroform solutions lead to sample contamination. Furthermore, chloroform leakage from the pipet tip can cause significant volume, and hence concentration, differences. Therefore, OMS concentrations were determined by carefully weighing the amount added to 161 Chapter 4.1 the sample. The final sample was placed into a standard 5 mm NMR-tube (Wilmad Royal Imperial). NMR SPECTROSCOPY All high-resolution 1H-NMR spectra were carried out at 298 K on a Bruker DRX 500 MHz spectrometer with a triple-resonance inverse (TXI) 1H {15N, 13C} probehead and equipped with x, y, z gradient coils. Shimming of the samples was performed automatically on the deuterium signal. The resonance linewidth for OMS was < 1 Hz in all spectra. For the one-dimensional spectra, 64 transients were recorded into 32000 data points using a spectral width of 6010 Hz and a 6 second recycle delay. A pulse width of 5 µs was used (corresponded to a 90° excitation pulse). An inversion recovery experiment revealed a T1 of 2.6 s for OMS. Data were processed and analyzed using MestReC version 4.4.1 (www.mestrec.com). The Free Induction Decay was apodized with a sine square filter and subsequently Fourier transformed after zero filling to 64 k points. The phase was corrected manually and metabolite signals were integrated (for peaks showing complex J-splitting) or fitted semi-automatically with a Lorentzian line shape (singlets only). The resulting areas were compared with the area of OMS to determine metabolite concentrations. Two-dimensional 1H-1H correlation (COSY) spectra were recorded using a spectral width of 6010 Hz in both dimensions, 256 and 2000 data points in F1 and F2, respectively, 16 scans per increment and a recycle delay of 6 seconds. Prior to Fourier transformation, both time domains were apodized using a sine bell function and zero filled once. Resonance assignments were based on literature [245;251-253], one and two-dimensional spectra of the authentic standards in chloroform and computer simulations using ACD/HNMR Predictor version 2.03 (ACD/labs, Toronto, Canada). Furthermore, available authentic standards of the accumulating metabolites were added to patient samples to confirm their assignment. Chemical shift values in this study are referenced to the chemical shift of OMS at 0.094 ppm. GAS CHROMATOGRAPHY For gas chromatography of sterols samples were extracted in pentane and subsequently derivatized using N,O-bis(trimethylsilyl)trifluoroacetamide (Fluka) and pyridine. GC analysis was performed using a CP-Sil-19 CB column (Chrompack, Bergen op Zoom, The Netherlands) on a Hewlett-Packard 6890 GC equipped with a flame ionization detector [254]. In the conventional analysis of plasma phytanic acid, all lipids were first extracted in 1:1 (v/v) chloroform-methanol. After evaporation of the organic solvent, the fatty acids were esterified using methanolic HCl. The resulting fatty acid methyl esters were subsequently extracted in hexane and analyzed with gas chromatography using the same equipment as described above for sterol analysis. 162 Inborn errors of lipid metabolism CORRELATION STUDY For the correlation study, 15 patient plasma samples were selected from routine samples of the clinical chemistry department to give a wide range of cholesterol and triglyceride values. They were selected anonymously and the results were not available to the person (M.O.) doing the 1H-NMR analysis. Cholesterol and triglycerides had been measured enzymatically in 15 and 12 samples, respectively, using standard reagent kits on an AEROSET System (Abbott laboratories, Illinois) according to instructions of the manufacturer [255;256]. For 1H-NMR analysis, samples were prepared as described above. Statistical analysis was performed using Passing and Bablok regression analysis [257]. The within run coefficient of variation of the NMR method was determined by preparing 9 separate samples from the same blood serum and subsequently measuring their cholesterol and triglyceride content. Sample preparation and NMR measurements were carried out in one session. INBORN ERRORS OF LIPID METABOLISM The patient population consisted of: • 3 cases of the Smith-Lemli-Opitz syndrome (SLOS) • 3 cases of cerebrotendinous xanthomatosis (CTX) • 1 case of sitosterolemia • 1 case of the Refsum disease. The patient material from the SLOS, CTX and sitosterolemia patients was obtained prior to treatment and therefore represents diagnostic samples. Six samples were obtained from the follow-up during therapy of a patient with Refsum disease. For every inborn error, previous detection of the accumulating metabolite as well as mutation analysis in the relevant gene had confirmed the diagnosis. The diseases are shortly described below. SLOS. The Smith-Lemli-Opitz syndrome is caused by a deficiency of the enzyme 7-dehydrocholesterol reductase (EC 1.3.1.21), which catalyzes the conversion of 7-dehydrocholesterol (7DHC, Figure 1) to cholesterol; the final step in cholesterol biosynthesis [258]. As a result, 7DHC accumulation and low cholesterol levels are observed in SLOS patients. Furthermore, 8-dehydrocholesterol (8DHC, Figure 1) can be detected in the blood of affected patients, resulting from isomerization of 7DHC to 8DHC. CTX. CTX is caused by a defect in the CYP27A gene, which encodes the mitochondrial enzyme sterol 27-hydroxylase (EC 1.14.13.15) [259]. This leads to a block in bile acid synthesis, resulting in the accumulation of unusual bile alcohols in urine and cholestanol in blood. 163 Chapter 4.1 Sitosterolemia. Sitosterolemia, is characterized by elevated levels of the plant sterols β-sitosterol, campesterol and stigmasterol (Figure 1) in blood and tissues. Mutations in the ABCG5 and ABCG8 genes, which both encode for half-transporter proteins, have been identified in sitosterolemia patients [260]. In healthy subjects, approximately 5% of the 200-300 mg daily consumed plant sterols is absorbed and almost all plant sterols are rapidly excreted in the bile. Sitosterolemia patients absorb between 15 and 60% of the ingested plant sterols and excrete only very little [259-261]. The mean β-sitosterol level in patients is approximately hundred fold increased [259]. Refsum disease. In Refsum disease patients, phytanic acid α-oxidation can not take place due to a defect in the enzyme phytanoyl-CoA hydroxylase (EC 1.14.11.18) [262]. Phytanic acid is totally exogenous in origin. It derives from the bacterial metabolism of chlorophyll in ruminants. Phytanic acid (3,7,11,15tetramethylhexadecanoic acid, Figure 1) accumulates in blood and tissues of affected patients [263]. Phytanic acid levels may reach values up to 1300 µmol/l (normal <10 µmol/l) in plasma, where it is incorporated in triglycerides [264]. Refsum patients are generally treated with a low phytanic acid diet [264]. Furthermore, plasmapheresis treatment can be used to rapidly lower plasma phytanic acid levels. Figure 1 Structures of the metabolites accumulating in the investigated inborn errors of lipid metabolism, including relevant atom numbering. For the plant sterols (g-i), only the side chain is shown. Their ring structure is identical to cholesterol. a cholesterol; b cholestanol; c 7DHC; d 8DHC; e 7lathosterol; f 8-lathosterol; g β-sitosterol; campesterol; i stigmasterol; j phytanic acid. 164 h Inborn errors of lipid metabolism Results NORMAL LIPID RESONANCES The one-dimensional 1H-NMR spectrum of the blood plasma lipid extract of a healthy volunteer is shown in Figure 1. Resonance assignments were based on previously reported spectra [245;251-253], one and two-dimensional spectra of pure cholesterol, oleic acid, linoleic acid, palmitic acid and stearic acid dissolved in deuterated chloroform and supporting computer simulations. In total, 25 lipid related resonances could be assigned, leading to the identification of nine molecular species, i.e. free cholesterol, esterified cholesterol, 7-lathosterol, triglycerides, spingomyelin, choline, glycerophospholipid, phosphatidylcholine and fatty acids (esterified plus non-esterified; Table 1). Although the spectrum in Figure 2 shows many overlapping signals from fatty acyl chains, several distinct groups can be identified, such as methyl, methylene, allylic (=CHCH2–), olefinic (–CH=CH–) and diallylic (=CHCH2CH=) protons (Table 1). Since these groups resonate at approximately the same frequency for every fatty acid, it is impossible to differentiate between the different fatty acids present in the sample. However, the use of methylene/diallylic and methylene/olefinic intensity ratio as an estimate for the average fatty acid chain length and number of double bonds, respectively, has been reported [251]. On the contrary, the C18H3 protons of cholesterol and its precursor 7-lathosterol are nicely resolved (See Figure 1 for molecular structures and atom numbering) and since the cholesterol C19 H3 group shows separate singlets for free and esterified cholesterol, it is possible to determine their relative amounts (approximately 60-70% of the total cholesterol is normally esterified). The C3 proton shows individual signals for esterified and nonesterified cholesterol as well, but unfortunately, the latter partially overlaps with the methanol contamination in Figure 2. Additionally, the C21 H3 and C26 H3/C27 H3 proton resonances can be distinguished, although they moderately overlap with fatty acid methyl groups. Finally, cross peaks between the allylic, olefinic and diallylic protons can be readily observed in a two-dimensional 1H-1H COSY spectrum of a healthy control, as well as cross peaks between the different protons of the glycerol backbone (data not shown). The crowded spectral region between 0.5 and 2.5 ppm could not be completely assigned due to overlap problems and complex J-splitting of the high number of cross peaks. However, changes in the observed cross peak pattern can still contribute to the diagnosis of an inherited metabolic disease (see below). 165 Chapter 4.1 Table 1: 1H resonance assignments with chemical shifts, multiplicity, and J-coupling constants for signals identified in the lipid extract of blood plasma taken from healthy controls (n = 3)a. Peak number 1 1H shift (ppm) 0.53 2 0.68 3 0.86 / 0.87 5 7 4 6 Multiplicity b ; J (Hz) Assignment Total 7-lathosterol C18H3 s Total cholesterol C18H3 s Total cholesterol C26H3 / C27H3 2x d; 6.6 0.91 Total cholesterol C21H3 d; 6.5 1.02 Esterified cholesterol C19H3 0.88 1.01 Fatty acyl chain CH3(CH2)n t; 6.9 Free cholesterol C19H3 s s c 8 1.05 – 1.19 Multiple cholesterol protons m 9 1.24 – 1.37 Fatty acyl chain (CH2)n m 10 1.42 – 1.55 Multiple cholesterol protons m 11 1.55 – 1.65 Fatty acyl chain –CH2CH2CO m 12 1.79 – 1.88 Multiple cholesterol protons m 13 1.98 – 2.09 Fatty acyl chain –CH2CH= m 15 16 2.77 – 2.87 3.32 / 3.35 Fatty acyl chain =CHCH2CH= 17 3.48 – 3.57 Free cholesterol C3H 19 3.96 14 18 20 21 22 2.24 – 2.35 3.81 4.15 / 4.29 4.32 – 4.43 4.57 – 4.65 23 5.17 – 5.24 20 5.26 21 5.29 – 5.43 Fatty acyl chain –CH2CO m Sphingomyelin and choline m N(CH3)3 d 2x s (largely overlapping) m Phosphatidylcholine N-CH2 d Glycerophospholipid backbone C3H2 Glycerol backbone C1H2 / C3H2 c Phosphatidylcholine PO-CH2 d Esterified cholesterol C3H c Glycerophospholipid backbone C2H d Glycerol backbone C2H c Fatty acyl chain –HC=CH– s (broad) d s (broad) m (AB spin system) m (broad) m m (broad) p; 5.7 m a; Peak numbers correspond with Fig. 1. Chemical shifts are referenced to OMS (δ = 0.094 ppm). All assignments were based on spectra of authentic standards, on the literature, or on computer simulations, unless specified differently. b; s, singlet; d, doublet; t, triplet; p, pentet; m, multiplet. c; Assignment based on the literature and spectral simulation. d; Assignment based solely on the literature. 166 tube-cleaning procedures, respectively. The peak labeled 13C corresponds to the 13C satellite of peak 2. The chemical shift reference compound OMS resonates at 0.094 ppm. The inset shows an expanded scale for the 0.52–0.58 ppm region. Peak assignment is provided in Table 1. Methanol (MeOH) and ethanol (EtOH) contamination resulted from extraction and 1-Dimensional 1H-NMR spectrum of the blood plasma lipid extract of a healthy female volunteer. Figure 2 Chapter 4.1 CORRELATION STUDY Cholesterol and triglyceride concentrations were measured in 15 and 12 blood plasma samples, respectively, using conventional enzymatic analysis and 1H-NMR spectroscopy. In the latter method, peaks 2 and 20 of Figure 2 were used for cholesterol and triglyceride quantification, respectively. The obtained results were compared using Passing and Bablok regression analysis [257], which showed a good correlation for both (Figure 3). Figure 3 Comparison concentrations conventional of a in cholesterol blood enzymatic and plasma analysis b triglyceride determined and by 1H-NMR spectroscopy. Solid lines correspond to the regression lines calculated with Passing and Bablok regression, while dashed lines indicate the 95% confidence intervals. For cholesterol, the regression line had a slope of 0.90 (95% confidence interval 0.83 – 1.07) and an intercept of 0.15 (-0.66 – 0.48) mmol/l, while for the triglycerides a slope of 1.02 (0.88 – 1.13) and an intercept of -0.04 (-0.31 – 0.10) mmol/l were obtained. No significant deviation from linearity (P>0.10) was found for both metabolites. However, it is clear that there is a slightly increasing deviation with increasing cholesterol concentration (Figure 3). This is likely due to a less efficient extraction at concentrations exceeding ~6 mmol/l and may be overcome by using a larger volume of 2:1 (v/v) chloroform-methanol. Another explanation might be the effect of a partial saturation of OMS. As the recycle time is not 5 times the T1-relaxtion time, OMS will be partially saturated. However, the methyl singlets of cholesterol and sterols have a similar T1-relaxtion as OMS (around 3 s). Therefore we estimate that the error made by partial saturation cannot be very significant. 168 Inborn errors of lipid metabolism The within run coefficient of variation was determined using 9 samples, which were prepared from the same blood serum containing 7.6 mmol/l cholesterol and 1.4 mmol/l triglyceride (values determined by NMR). For cholesterol and triglyceride concentrations, a within run coefficient of variation of 7.9% and 6.8% was found, respectively. The detection limit may vary for different metabolites, since it is dependent on the number of equivalent protons contributing to the NMR signal, the peak splitting pattern, the number of scans and the field strength of the spectrometer. The detection limit for the C18 H3 singlet of cholesterol at 0.68 ppm is estimated to be approximately 10 µmol/l, assuming that the peak can be distinguished when the signal to noise ratio is ≥3. INBORN ERRORS OF METABOLISM To assess the diagnostic ability of 1H-NMR spectroscopy of lipid extracts, several samples from patients with a known inborn error of lipid metabolism were investigated. Three diseases involved errors in sterol metabolism, while one disease is caused by a defect in the breakdown of an unusual branched chain fatty acid. Table 2 gives characteristic resonance frequencies of different metabolites. Table 2: Characteristic resonance assignments, including multiplicity and J-coupling constants (Hz) used for identification of metabolites accumulating in several inborn errors of lipid metabolism1. Metabolite Assignment 1H shift (ppm) Multiplicity2 (J-coupling) Inborn error 7DHC C-18 H3 0.618 s SLOS Cholestanol C-18 H3 0.645 s CTX 7-Lathosterol C-18 H3 0.534 s CTX, lathosterolosis4 8DHC3 C-18 H3 C-19 H3 8-Lathosterol3 ß -Sitosterol s 0.817 C-19 H3 0.810 C-18 H3 0.680 C-18 H3 s s 0.606 s C-26 H3/C-27 H3/C-29 H3 0.800–0.850 m C-24 H 1.685 Phytanic acid (esterified to glycerol) C-16 H3/C-17 H3 C-18 H3/C-19 H3 C-20 H3 1 0.651 m 0.844 d (6.6) 0.867 0.9273, s SLOS CTX Sitosterolemia Refsum disease d (6.6) 5 d (5.6) Chemical shifts are referenced to OMS (δ = 0.094 ppm). Values were determined in patients and confirmed by use of authentic standards, unless stated otherwise. 2 3 4 5 s, singlet; d, doublet; m, multiplet. Tentative assignment; no authentic standard available. No patient material available. Resonates at higher ppm value in pure compound (see text). 169 Chapter 4.1 Smith-Lemli-Opitz syndrome (SLOS; OMIM 270400) Figure 4b shows the 0.55 – 0.75 ppm region of the 1D 1H-NMR spectrum of the blood plasma lipid extract of a 5 months old boy suffering from SLOS. It clearly differs from the spectrum of a healthy volunteer (Figure 4a) and shows the diagnostic metabolites 7DHC and 8DHC. The characteristic C18 H3 resonances used for identification and quantification of cholesterol, 7DHC and 8DHC are nicely resolved. The presence of 7DHC was confirmed by addition of the pure compound to the sample and remeasuring the NMR spectrum. An addition experiment could not be performed for 8DHC because this compound is not commercially available. More certain identification of both metabolites can be achieved when additional signals of these compounds are resolved. Unfortunately, the current experimental conditions lack the required higher resolution. Samples from two other SLOS patients showed a similar NMR spectrum. Ruan et al. [248] and Xiong et al. [249], assigned 7DHC and 8DHC unequivocally in an 1HNMR spectrum. Our study for the first time quantifies both metabolites. The quantitative data for 7DHC and 8DHC show discrepancies between NMR and GC (Table 3). These may relate to the long time interval between GC and NMR measurements (in some cases several years). It is known that the unsaturated sterols decompose easily during storage, processing and/or analysis. Accumulation of cholesta-5,7,9(11)-triene-3β-ol in the blood of SLOS patients has been reported [248]. The C18 H3 protons of this compound has been reported to resonate at 0.566 ppm [245]. However, we observed no resonance at this position in the NMR spectra of SLOS patients. The concentration of this compound varies considerably and may be in the low micromolar range in SLOS patients. Ruan et al. reported concentrations from 0.8 to 79.4 µmol/l [248]. This may explain why we were unable to detect this compound in our patient samples. 170 Inborn errors of lipid metabolism Figure 4 Upfield regions of the 1-dimensional 1H-NMR spectra of patient plasma lipid extracts, plasma from a healthy control, and the authentic ß-sitosterol and phytanic acid standards. Methyl assignment for both authentic standards are shown (see Table 2 atom numbering). (A), control; (B), SLOS; (C), CTX; (D), control; (E), sitosterolemia; (F), ß-sitosterol; (G), Refsum disease; (H), phytanic acid. Peaks in panels A–C correspond to the C- 18 H3 resonance of the indicated metabolite. The peak labeled 13C in panel C represents the cholesterol C-18 H3 13C satellite, whereas the peak with unknown assignment is labeled with an X. The peak numbering in panel D is identical to the numbering in Fig. 1 . Peaks labeled with an * in panels E and G correspond to ß-sitosterol (sito) and phytanic acid signals, respectively. The C-20 H3 resonance of esterified phytanic acid is indicated by # in panel G. Chol (panel E), cholesterol. Cerebrotendinous xanthomatosis (CTX; OMIM 213700) The 1D lipid 1H-NMR spectrum of a 24-year-old male CTX patient is given in Figure 4c. It shows an abnormally high concentration of the diagnostic metabolite cholestanol (C18 H3 resonance at 0.645 ppm, close to the C18 H3 of cholesterol). The NMR method revealed a cholestanol concentration of 0.15 mmol/l, which is somewhat higher than the value determined by GC (0.11 mmol/l, Table 3). This may be due to slight overlap with the tail of the cholesterol resonance. Furthermore, the spectrum shows three other peaks: the 13C satellite of the cholesterol C18 H3, the C18 H3 resonance of 7-lathosterol and one unknown 171 Chapter 4.1 metabolite (peak X). Although 7-lathosterol is also observed in healthy controls (Figure 2), elevated levels were found in all investigated CTX patients (Table 3), which corresponds with reference values determined by Wolthers et al. [265]. Interestingly, peak X at 0.61 ppm does not occur in control spectra, but was present in the two other CTX patients as well (both female and before start of therapy; 16 and 46 years old), although it was much weaker in the 16 year old female. GC and GC-MS measurements also displayed an unknown peak in CTX patients, which was tentatively identified as 8-lathosterol (Figure 2). Based on this finding and substantial supporting evidence, peak X is tentatively assigned as the C18 H3 resonance of 8-lathosterol. First, the relative peak area compared to cholestanol was very similar in the gas chromatogram and the 1D 1H-NMR spectrum of the same patient. Second, the lathosterol isomers show the same structural difference as 7DHC compared to 8DHC, i.e. the position of the double bond (Figure 2), and the frequency of peak X differs from 7-lathosterol in the same way as 8DHC from 7DHC (cf. Figure 4b). Although the chemical shift difference between the 7-lathosterol and 8-lathosterol and the chemical shift difference between 7DHC and 8DHC is not identical (0.073 and 0.031 ppm respectively), this does provide some tentative evidence. Third, our assignment is strongly supported by corresponding chemical shift values found by Wilson et al. for 7- and 8-lathosterol [245]. Finally, results described by Wolthers et al. confirm the presence of 8-lathosterol in plasma of CTX patients [265] and the reported concentration (60.0 µmol/l; n=1) corresponds very well with the average concentration of 66.8 µmol/l determined by 1H-NMR (Table 3). Unfortunately, 8lathosterol is not commercially available for absolute confirmation of the assignment. Furthermore, the current resolution does not allow identification of additional 8-lathosterol signals. Sitosterolemia (OMIM 210250) Figure 4e shows the upfield region of the one-dimensional 1H-NMR spectrum of the lipid extract of a sitosterolemic patient. For comparison, the same spectral region is also shown for a healthy control and pure β-sitosterol in Figures 4d and f, respectively. In the analysis, only β-sitosterol was taken into account, since this is the major plant sterol, constituting about 65% of all absorbed plant sterols [259]. Campesterol and stigmasterol contribute approximately 32% and 3%, respectively. Unfortunately, β-sitosterol quantification was impossible under the current experimental conditions, since the C18 H3 and C19 H3 resonances of β sitosterol and cholesterol overlap almost completely (Figure 4e) and other βsitosterol resonances were not resolved into individual doublets and triplets to allow accurate quantification. Although the assignment of the C26 H3, C27 H3 and C29 H3 resonances (0.80 – 0.85 ppm) is therefore not exactly known, their presence in the sitosterolemia patient and absence in the healthy control is clear. Peak positions in the authentic standard corresponded exactly with the observed positions in the patient and repeated NMR measurement after addition of pure β - 172 Inborn errors of lipid metabolism sitosterol confirmed the assignment. The weak resonances at approximately 0.75 to 0.80 ppm (at right of the β-sitosterol signals) likely result from campesterol and dihydrobrassicasterol (24S stereo-isomer of campesterol), since corresponding peak positions were found in the authentic standard of campesterol (data not shown). Overlap with the campesterol signals provides an additional difficulty in quantifying β-sitosterol. Therefore, higher resolution and precise knowledge of all assignments is necessary to allow quantification of both β-sitosterol and campesterol. To support the diagnosis, a two-dimensional 1H-1H COSY spectrum can be recorded. The resulting spectrum shows a characteristic cross peak at (1.68; 0.82) ppm, originating from the long-range spin-spin coupling between C24 H and C26 H3, C27 H3 or C29 H3 of β-sitosterol (data not shown). Thus, β-sitosterol presence can be measured using our NMR method, although quantification is still impossible. Figure 5 Effect of plasmapheresis treatment on plasma phytanic acid concentrations in a female Refsum disease patient. The duration of the plasmapheresis is indicated in gray. Refsum disease (OMIM 266500) Figure 4g shows the 0.60 – 1.05 ppm region of the 1D 1H lipid NMR spectrum of a 29 year old female Refsum disease patient. Assignment of the methyl resonances is given in the spectrum of pure phytanic acid (Figure 4h). The doublet at 0.84/0.85 ppm provides clear evidence for the presence of phytanic acid in the Refsum sample (cf. Figure 4g). Remarkably, the C20 H3 protons, which resonate as a doublet at 0.97/0.98 ppm in pure phytanic acid, appear absent in the patient spectrum. However, due to esterification with glycerol [264], this peak has probably shifted to 0.92/0.93 ppm, where a clear difference is seen between the Refsum case and a healthy control. After addition of pure phytanic acid to the patient sample, the doublet at 0.97/0.98 ppm was evident, confirming that phytanic acid is not present in its free form in Refsum disease patients. Additional computer simulations confirmed a shift towards lower ppm values upon esterification with glycerol. 173 Chapter 4.1 Phytanic acid quantification by 1H-NMR (using half of the C16 H3/C17 H3 doublet at 0.84 ppm) showed good correspondence with values determined by GC (Table 3). The effect of plasmapheresis on the phytanic acid concentration was followed using NMR analysis. A rapid decrease in phytanic acid concentration of 50% was observed. A return to its original value was found three days after the treatment (Figure 5). Table 3: Comparison of plasma lipid concentrations by 1H-NMR and gas chromatography for the Smith-Lemli-Opitz syndrome, cerebrotendinous xanthomatosis and Refsum disease. All values are reported in µmol/l, except for cholesterol, which is given in mmol/l. Reference values (Ref) were taken from Blau et al. [140] , and for 7-lathosterol from Wolthers et al. [265] . No reference value was available for 8-lathosterol. Smith-Lemli-Opitz syndrome 7DHC Cholesterol NMR GC Ref NMR 8DHC GC Ref NMR GC Ref 1 (M, 5 m) 2.12 2.05 2.96-4.43 332 334 <0.02-0.57 363 225 <0.02 2 (M, 8 m) 0.39 0.26 2.96-4.43 153 353 <0.02-0.57 143 385 <0.02 3 (M, 7 y) 1.25 1.10 2.89-4.74 200 325 0.03-0.52 232 295 <0.02 Cerebrotendinous xanthomatosis Cholesterol Cholestanol 7-lathosterol NMR GC Ref NMR GC Ref NMR GC 1 (M, 24 y) 3.10 3.00 2.66-6.02 152 106 4-18 34 ND 2 (F, 16 y) 4.72 4.20 2.66-6.02 138 101 4-18 11 3 (F, 46 y) 4.30 3.90 2.66-6.02 127 125 4-18 39 Ref NMR 8-lathosterol Ref NMR GC <10.6 75 ND ND <10.6 7 ND ND <10.6 118 ND a b Refsum disease Cholesterol 1 (F, 29 y) NMR 1.90 GC ND Phytanic acid 2.66-6.02 GC Ref 700 690 <10 2 (F, 29 y) 504 454 <10 3 (F, 29 y) 354 278 <10 4 (F, 29 y) 546 528 <10 5 (F, 29 y) 700 588 <10 6 (F, 29 y) 722 658 <10 ND: not determined b: tentative assignment a Discussion Many inherited metabolic diseases are caused by defects in lipid metabolism or biosynthesis, often resulting in accumulation of unusual lipids in the blood and tissues of affected patients. In this study, the successful use of 1H-NMR spectroscopy of blood plasma lipid extracts to accurately identify and quantify these lipids was demonstrated for four different inborn errors of lipid metabolism. 174 Inborn errors of lipid metabolism Twenty-five lipid derived resonances could be assigned in the 1D spectra of healthy controls, resulting in the identification of nine distinct molecular species. This number increases to fourteen when also taking the investigated metabolic diseases into account. Although 1H-NMR spectroscopy of lipid extracts was used previously by Kriat et al. [251] to study tumor induced effects, its application in clinical diagnosis has hitherto been only described for the identification of unusual metabolites in the blood plasma of Smith-Lemli-Opitz syndrome patients [248;249]. By using OMS instead of the highly volatile TMS as a concentration and chemical shift reference, we are for the first time able to accurately quantify metabolites in chloroform solutions by proton NMR. A good correlation with conventional methods was obtained for total cholesterol and triglyceride concentrations. This does not exclude a possible less efficient extraction and hence worse correlation for other metabolites. Based on the results obtained for cholesterol, a larger extraction volume has to be used when high lipid concentrations are expected. Nevertheless, the currently applied protocol will likely be sufficient for the quantification of unusual lipids since their concentrations are normally low (<1 mmol/l) and deviations are only seen at high concentrations (>6 mmol/l for cholesterol). Since lipid 1H-NMR spectroscopy requires approximately one day of sample preparation and only a limited number of samples can be worked up, it can not compete with the automated enzymatic analyses of cholesterol and triglycerides in blood. Automated extraction procedure may be a further solution. However, it does provide an excellent alternative for the identification and quantification of unusual lipids. For instance, cholestanol and cholesterol separation can be tedious using conventional chromatographic methods, whereas both metabolites are readily identified with 1H-NMR spectroscopy based on their different C18 H3 resonance. Similarly, the presence of 7DHC and 8DHC in the plasma of SLOS patients is immediately evident from a simple 1D 1H-NMR spectrum. Quantification of 8DHC may be more accurate using 1H-NMR, since chromatographic measurements can not be calibrated correctly as the authentic standard is not commercially available. The diagnosis of SLOS, CTX and sitosterolemia can be made with almost a 100% certainty based on NMR analysis. However, for Refsum disease, the total clinical picture is required, since phytanic acid accumulation is also observed in the Zellweger syndrome, neonatal adrenoleukodystrophy, infantile Refsum disease and rhizomelic chondrodysplasia punctata type 1 [264]. The investigation of these diseases using 1H lipid NMR spectroscopy is a topic for future research. A database containing the resonances of many more authentic standards of molecules relevant in lipid metabolism has to be established, which will be helpful to assign unknown signals. For NMR spectra of urine samples, an analogous strategy has led to the identification of new inborn errors of metabolism [33;36;42] and it is not unlikely that lipid NMR spectroscopy can also lead to the discovery of currently unknown inherited disorders. 175 Chapter 4.1 Besides resonance positions, which can only provide clear information when there is no overlap, peak ratios can be used to obtain additional important information. For example, the CH2:CH3 ratio (peaks 9 and 4 in Figure 2, respectively) can provide evidence on the presence of branched chain fatty acids since this will lower this ratio. For healthy individuals (n=3), we found a CH2:CH3 peak ratio of 6.61 ± 0.07 (note that the CH3 integral has to be corrected for overlap with cholesterol resonances). This corresponds very well with a ratio of 6.67 calculated for a theoretical 1:1:1 mixture of palmitic, oleic and linoleic acid, which together constitute approximately 75% of all plasma fatty acids. As expected, a much lower ratio of 3.87 was found for our Refsum disease patient, which is fully explained by the branching of the phytanic acid and its concentration compared to other fatty acids in the sample. Although the presence of phytanic acid can be more easily ascertained based on detection of the C16 H3/C17 H3 resonance (Figure 1), it is not unlikely that peak ratios can be helpful to detect abnormalities in spectra without any clear unusual signals. In conclusion, we have presented a new method to identify (un)usual lipids in blood plasma or serum using 1H-NMR spectroscopy. Furthermore, metabolite concentrations can be accurately determined by using the non-volatile OMS as a chemical shift and concentration reference compound. The technique is well suited for the diagnosis and follow up of inborn errors of lipid metabolism, which was demonstrated for four different inherited diseases. Acknowledgements All 1H-NMR spectra were recorded at the Dutch hf-NMR facility at the Department of Biophysical Chemistry, Radboud University Nijmegen, The Netherlands (head Prof. Dr. S.S. Wijmenga). We thank Jos Joordens for his technical assistance and Fokje Zijlstra and Arno van Rooy for sterol analysis with conventional GC and GCMS techniques. We also thank the department of clinical chemistry (head Prof. Dr. J.L. Willems) for sample selection and for the enzymatic determination of cholesterol and triglyceride concentrations. Finally, we thank Prof. Dr. A.F.H. Stalenhoef, Dr. E. Morava, Dr. G. Drost, Dr. A. Verrips, Dr. P. Jira and Dr. L.J.H. van Tits for referring relevant patient samples. 176 FIVE SUMMARY AND FUTURE PERSPECTIVES 177 178 CHAPTER 5.1 Summary 179 180 Summary Summary This thesis describes the use of proton nuclear magnetic resonance (1H-NMR) spectroscopy as a tool for diagnosing inborn errors of metabolism in body fluids. The technique yields an overview of proton-containing metabolites, is rapid and nondestructive, and requires minimal sample pretreatment. Moreover, structural information from metabolites can be obtained from NMR spectra. Our studies demonstrate that NMR spectroscopy of body fluids may be considered as an alternative analytical approach for diagnosing known, but also as yet unknown, inborn errors of metabolism. In this thesis we describe the NMR spectroscopy contributions to the unraveling of two novel inborn errors of metabolism. Chapter 1 The aim of our study is described in chapter 1.1 and current knowledge about NMR spectroscopy in the field of inborn errors of metabolism is reviewed in chapter 1.3. Chapter 2 The second chapter describes the NMR findings for a group of compounds that normally will not be detected in a metabolic laboratory. These compounds contain an N-acetyl group (Figure 1). Figure 1 General structure of an N-acetylated compound. The three equivalent protons from thr N-acetyl group resonate as a singlet in a small part of the 1H-NMR spectrum. In chapter 2.1 13 inborn errors of metabolism that involve N-acetylated compounds are studied using 1H-NMR spectroscopy. Urine samples were measured from patients suffering from Salla disease, French-type sialuria, αmannosidosis, β-mannosidosis, fucosidosis, aspartylglucosaminuria, GM1gangliosidosis, GM2-gangliosidosis, sialidosis, Canavan disease, citrullinemia, tyrosinemia type I and tyrosinemia type II. Diagnostic important N-acetylated metabolites could be detected in the urine 1D and 2D NMR spectra of patients of almost all inborn errors that were measured. No abnormalities were observed in the spectra of urine from patients with GM1- or GM2-gangliosidosis. Chapters 2.2 and 2.3 demonstrate the diagnostic importance of N-acetylated metabolites. In 181 Chapter 5.1 Chapter 2.2, in vivo and in vitro NMR spectroscopic investigations are described in two patients. The patients are two unrelated girls with almost absent CNS myelin, intractable epilepsy, nystagmus, progressive microcephaly, and severe psychomotor retardation. Moreover, in the 1H-NMR spectra of cerebrospinal fluid of the patients, high concentrations of N-acetylaspartylglutamate (NAAG) could be observed. Elevated NAAG concentration in CSF may be the biochemical hallmark for a novel neurometabolic disorder and may provide a key for understanding the molecular defect of this disease. Chapter 2.3 describes the in vitro NMR spectroscopic investigations at the metabolite level on eight unrelated patients with aminoacylase 1 deficiency; a novel disease whereby NMR spectroscopy was involved in the elucidation of the molecular defect. The 1D 1H-NMR urine spectra from these patients show a characteristic profile, mainly due to the N-acetyl singlets deriving from N-acetylated amino acids. Using a combination of several 1D and 2D NMR experiments, resonances of eleven N-acetylated amino acids could be assigned and quantified. Chapter 3 In this chapter we used both in vitro and in vitro NMR spectroscopic techniques to identify unusually high resonances in the brain and body fluids. In chapter 3.1, several NMR experiments are described to identify an unknown peak in NMR spectra of human cerebrospinal fluid and blood plasma. The peak was assigned to dimethyl sulfone (DMSO2). In plasma and CSF from healthy controls, the concentration of DMSO2 ranged between 0 and 25 µmol/l. For the first time we desribe the presence of DMSO2 human CSF. Furthermore, we document an increased concentration of DMSO2 in plasma from three out of four patients with methionine adenosyltransferase I/III (MAT I/III) deficiency. Medication, food or ‘innocent’ supplements can cause unusual resonances in NMR spectra that may be misinterpreted as diagnostic resonances. Chapter 3.2 reports the detection of DMSO2 (also known as methylsulfonylmethane (MSM) or sulfonylbismethane) in the brain and CSF of a 4-year-old girl using in vivo and in vitro NMR spectroscopy. The presence of this exogenous metabolite resulted from ingestion of a dietary supplement containing DMSO2. The concentration of this compound amounted to 1.2 mmol/l in brain tissue and 1.7 mmol/l in cerebrospinal fluid. Chapter 3.3 describes in vivo and in vitro NMR spectroscopic investigations on a 61-year-old woman with a 3-methylglutaconyl-CoA hydratase (3MGH) deficiency or 3-methylglutaconic aciduria type I. MR imaging of her brain revealed a leukoencephalopathy. Increased concentrations of 3-hydroxyisovaleric acid (3HIVA), 3-methylglutaconic acid (3MGA) (cis and trans) and 3-methylglutaric acid (3MG) were observed in the NMR spectra of the patient’s urine. In the CSF, the 3HIVA concentration was 10 times higher than in the plasma of the patient and only the cis isomer of 3MGA 182 Summary was observed. In vivo brain MRSI showed an abnormal resonance at 1.28 ppm that may be caused by 3HIVA. The brain 3HIVA was calculated to be 0.9 mmol/l, about five times the concentration in CSF. Furthermore, four other diseases involving leucine metabolism are described. Urine samples were measured from patients suffering from maple syrup urine disease, isovaleric academia, 3-methylcrotonyl CoA carboxylase deficiency and 3-hydroxy-3-methylglutaryl-CoA lyase deficiency. Diagnostic metabolites could be detected in the urine NMR spectra of all patients of these inborn errors of metabolism. Chapter 3.4 is a case report on the 61-yearold woman, described in chapter 3.3. This case demonstrates four features of 3methylglutaconic aciduria type I. The patient is unique in terms of age of presentation and perhaps of onset of her disease; 3-methylglutaconic aciduria type I may present as a leukoencephalopathy; in vivo MR spectroscopy may demonstrate increased brain tissue levels of 3HIVA. Furthermore we describe two novel apparently mild mutations. Chapter 4 Chapter 4 investigates the potential use of 1H-NMR spectroscopy to identify and quantify lipids present in the blood of patients with different inborn errors of lipid metabolism. We extracted blood plasma or serum lipids in chloroform–methanol (2:1 by volume). After addition of the nonvolatile chemical shift and concentration reference compound octamethyl-cyclotetrasiloxane (OMS), we performed 1H-NMR measurements. Blood plasma or serum lipid extracts were measured from controls and patients suffering from Smith–Lemli–Opitz syndrome, cerebrotendinous xanthomatosis, sitosterolemia, and Refsum disease. NMR analysis led to a correct diagnosis for all 4 diseases, whereas the concentration of the diagnostic metabolite could be determined for 3. Simultaneous lipid quantification was possible for the first time. 183 Chapter 5.1 Samenvatting De Nederlandse titel van mijn proefschrift is: NMR-spectroscopie van lichaams- vloeistoffen; een metabolomics benadering van erfelijke stofwisselingsziekten. In deze titel staan een aantal wetenschappelijke termen die voor de niet- wetenschapper mogelijk moeilijk te begrijpen zijn. In onderstaande tekst zal ik de titel en mijn onderzoek nader toelichten. Er bestaan meer dan vijfhonderd verschillende erfelijke stofwisselingsziekten die allemaal moeilijke namen hebben. Op pagina 44 in tabel 2 van dit proefschrift staan enkele voorbeelden van deze ziekten. Elke stofwisselingsziekte afzonderlijk is zeldzaam; gezamenlijk vormen ze een klinisch belangrijke categorie ziekten. Stofwisselingsziekten komen voor in alle leeftijdsgroepen, maar het meest bij kinderen. Vanaf 1 januari 2007 wordt het bloed van pasgeboren baby´s in Nederland onderzocht op de aanwezigheid van veertien stofwisselingsziekten (de hielprik). Vroege opsporing en behandeling van deze ziekten kan blijvende schade voorkomen. Wat is een stofwisselingsziekte? Bij patiënten met een erfelijke stofwisselingsziekte is de aanmaak of verwerking van één of meerdere stoffen (metabolieten) verstoord. Meestal wordt dit veroorzaakt doordat één onderdeeltje van alle cellen, een enzym, niet (goed) werkt. Het gevolg van die verstoring is dat er ergens in het lichaam een teveel of tekort aan een bepaalde metaboliet ontstaat. Ook in de lichaamsvloeistoffen (urine, bloed of hersenvloeistof) van deze patiënten kan een te laag of juist te hoog gehalte (concentratie) van een bepaalde metaboliet aanwezig zijn. Hoe kunnen we deze ziekten opsporen? Bij een afwijkende metabolietenconcentratie in een lichaamsvloeistof kan er dus sprake zijn van een stofwisselingsziekte. Het kan daarbij om te hoge of te lage concentraties van een bepaalde metaboliet gaan. Metabolomics is het opsporen en bepalen van concentraties van alle metabolieten in een cel of een lichaamsvloeistof. Een ziekenhuislaboratorium dat gespecialiseerd is in het diagnosticeren (opsporen) van erfelijke stofwisselingsziekten gebruikt verschillende technieken voor het routinematig bepalen van metabolieten in lichaamsvloeistoffen. Metabolieten kan men onderverdelen in groepen op basis van hun chemische eigenschappen. Zo zijn er bijvoorbeeld suikers, aminozuren of organische zuren. Voor de bepaling van de diverse groepen metabolieten heb je in een laboratorium verschillende technieken nodig. 184 Samenvatting In dit proefschrift gebruik ik een nieuwe analysetechniek die in tegenstelling tot de andere technieken vrijwel alle metabolieten tegelijk kan aantonen. Deze techniek noemen we NMR-spectroscopie. NMR-spectroscopie van lichaamsvloeistoffen Metabolieten hebben met elkaar gemeen dat ze allemaal waterstofatomen bevatten. In de scheikunde gebruiken we hiervoor de afkorting 1H. Met een sterke magneet (Figuur 1, bladzijde 19) en radiogolven kunnen we deze waterstofatomen zichtbaar maken; we noemen dit een 1H-NMR-experiment. Een urine-1H-NMRexperiment duurt ongeveer dertig minuten. Het resultaat van een 1H-NMR- experiment zijn patronen van pieken; dit noemen we een 1H-NMR-spectrum. Een 1H-NMR-spectrum van een urine van een gezonde persoon bestaat uit meer dan driehonderd pieken; elke piek wordt veroorzaakt door één bepaalde metaboliet (Figuur 4, bladzijde 41). Urine-1H-NMR-spectra van patiënten met een erfelijke stofwisselingsziekte tonen vaak een afwijkend piekenpatroon. Het kan daarbij gaan om de aanwezigheid van abnormaal hoge concentratie van een bepaalde stof of juist de afwezigheid van een bepaalde stof. Met een 1H-NMR-spectrum kunnen we dus iets zeggen over de mate van aan- of afwezigheid van bepaalde metabolieten in een lichaamsvloeistof. Deze informatie gebruiken we voor het opsporen van een stofwisselingsziekte. Wat symboliseert de omslag van dit boek? Op de omslag van mijn proefschrift staan ongeveer tweehonderd metabolieten die voorkomen in de mens. Al deze metabolieten kunnen we met spectroscopie eenvoudig meten. In 1H-NMR-spectra 1H-NMR- van patiënten met een stofwisselingsziekte zijn soms heel hoge concentraties van een bepaalde metaboliet te meten (een hoge piek in het NMR-spectrum). Soms is de hoeveelheid van een bepaalde metaboliet nauwelijks zichtbaar (een kleine piek in het NMRspectrum). In hoeverre we de concentratie van een bepaalde metaboliet kunnen vaststellen, verschilt dus per stofwisselingsziekte. Het proefschrift In dit proefschrift heb ik de mogelijkheden van NMR-spectroscopie beschreven om in lichaamsvloeistoffen erfelijke stofwisselingsziekten op te sporen. In hoofdstuk 2 heb ik één groep metabolieten onderzocht die binnen een normaal ziekenhuislaboratorium niet worden geanalyseerd (N-acetylmetabolieten). In dit hoofdstuk beschrijf ik de NMR-resultaten van dertien bekende erfelijke stofwisselingsziekten (hoofdstuk 2.1) en twee ziekten die nog niet eerder beschreven zijn (hoofdstuk 2.2 en 2.3). In de urine- of hersenvocht-1H-NMRspectra van deze patiënten werden steeds afwijkende pieken gevonden die 185 Chapter 5.1 veroorzaakt werden door één of meerdere N-acetylmetabolieten. Aan de hand van deze metabolieten konden we de betreffende ziekten vaststellen. Met NMR is het ook mogelijk om metabolieten te meten binnen organen van het menselijke lichaam. Bij deze experimenten wordt de patiënt in een grote speciale magneet geschoven. Met de magneet en radiogolven wordt eerst een beeld gemaakt van bijvoorbeeld de hersenen. Daar kunnen we van een klein volume een 1H-NMR-spectrum hoofdstuk 3 opnemen. We noemen dit in vivo MR-spectroscopie (MRS). In gebruiken we deze techniek in combinatie met 1H-NMR- spectroscopie van lichaamsvloeistoffen om ongewone pieken te identificeren. Hoofdstuk 3.1 en 3.2 beschrijft de ontdekking van de stof dimethyl sulfone in hersenvocht, bloed en de hersenen. Hoofdstuk 3.3 en 3.4 gaat over een patient met een zeldzame stofwisselingsziekte. In de 1H-NMR-spectra van de lichaamsvloeistoffen en de hersenen vonden we ongewone pieken. Deze konden we allemaal identificeren met NMR. Hoofdstuk 4 staat in het teken van lipiden. Zij vormen een speciale groep metabolieten. Lipiden zijn vetten en vetachtige stoffen die in de stofwisseling van de mens een belangrijke rol spelen. Enkele voorbeelden hiervan zijn cholesterol, fosfolipiden en triglyceriden. Het is bekend dat een aantal lipiden een rol spelen in erfelijke stofwisselingsziekten. In dit hoofdstuk hebben we de mogelijkheden van NMR-spectroscopie onderzocht om lipiden te meten in bloedplasma. Door bloedplasma op een speciale manier te behandelen konden we de lipiden zichtbaar maken in een 1H-NMR-spectrum. Bloed van gezonde vrijwilligers hebben we vergeleken met patiënten. Vier verschillende erfelijke stofwisselingsziekten werden onderzocht. Bekend was dat bij deze ziekten lipiden betrokken waren. NMR-spectroscopie gaf een correcte diagnose voor de vier ziekten. Dit is de eerste studie die laat zien dat het mogelijk is met NMR-spectroscopie diverse lipiden aan te tonen en te kwantificeren in biologische monsters. 186 CHAPTER 5.2 Future perspectives 187 Chapter 5.2 188 Future perspectives Future perspectives Although many inborn errors of metabolism can be readily diagnosed with NMR, the technique is not yet routinely available in the majority of hospital laboratories. This is likely due to the high cost of the equipment and the fact that spectrum interpretation is usually not straightforward, requiring highly trained personnel. Several ongoing developments, which may make the application of body fluid NMR in diagnosing inborn errors of metabolism more successful and hopefully also more accessible, will be discussed in this chapter. The first three paragraphs will deal with advances made in NMR hardware technology, while the final two paragraphs will cover some of the progress in spectral analysis software and pulse sequence design. DETECTION OF OTHER NUCLEI Protons are highly abundant in many metabolites and have a high sensitivity, making them the most widely used nuclei in the study of inborn errors of metabolism. However, several other NMR active nuclei, which can provide important additional information, are present as well. Although 19F and 31P have a favorable natural abundance and relative sensitivity, there are few phosphorous or fluorine-containing metabolites present in body fluids. On the contrary, 13C and 15N are present in many metabolites, but suffer from their low inherent sensitivity and natural abundance. These disadvantages may be overcome with the use of cryogenic probe technology, which was first presented by Styles et al. in 1984 [266]. By cooling the receiver coil and preamplifier to approximately 20 K with liquid helium, the resistance of the coil is significantly reduced. Consequently, the thermal noise is decreased by approximately a factor 4. This leads to a corresponding increase in the signal-to-noise ratio (S/N) per scan or, for the same S/N, a 16-fold reduction in acquisition time when compared to a normal probe. The improved sensitivity is such that 13C spectra with good S/N can be obtained for body fluid samples with acceptable acquisition times [267]. The higher sensitivity obtained by using cryoprobe technology can be advantageous for proton NMR as well. With the resulting decrease in detection limit it is likely that NMR can diagnose many additional inborn errors of metabolism, where relatively low amounts of metabolites accumulate. It furthermore may facilitate the discovery of new diseases. The low natural abundance of several nuclei can also be used as an advantage in metabolic investigations, since their weak background signal allows the study of the metabolism of labeled compounds. An example of this technique was already given in 1975 by Tanaka et al. [268], who elucidated the metabolic pathway of 13C-labelled valine in a boy suffering from methylmalonic acidemia by administering the patient orally with DL-[α-13C]-valine and DL-[α, β-13C]-valine. 189 Chapter 5.2 LINKING NMR WITH OTHER ANALYTICAL TECHNIQUES On one hand, NMR has the advantage that many metabolites can be detected simultaneously and that no preselection is required. On the other hand, this is also a disadvantage, since it can cause highly complex spectra with a considerable amount of overlap. By linking NMR to HPLC it is possible to separate and directly identify the compounds present in complex mixtures. The first example employing coupled 500 MHz HPLC-NMR was presented in 1992 by Spraul et al. [269], who investigated the metabolism of the drug ibuprofen in urine samples. Both stopped-flow and continuous-flow experiments were performed. In 1995, the first set up combining NMR with both HPLC and MS was presented [270] and the stopped-flow method with parallel NMR and MS detection has now become standard [271]. In body fluid research, HPLC-NMR and HPLC-NMR-MS have mainly been used to investigate the metabolic fate of different drugs and potential drug candidates [28;269;272-275]. Both methods have proven to be valuable tools to obtain structural and pharmaceutical data of metabolites present in complex matrices with a single experiment. In the field of inborn errors of metabolism, it is expected that the combination of NMR with liquid chromatography and/or mass spectrometry can also be helpful in elucidating (partially) unknown pathways. Furthermore, the detection of new metabolites and their associated diseases may become less complicated. MICROPROBES Over the years, NMR has undergone some major developments, but the conventional 5 mm sample tube used today is still essentially the same as in the 1960s. Historically, relatively large amounts of sample were needed to compensate for the low inherent NMR sensitivity. Nowadays, ultra-high-field spectrometers equipped with cryoprobe technology are becoming available, making measurement of mass-limited samples theoretically possible. The sensitivity of these samples can be further enhanced by using microprobes. With decreasing diameter d of the detector coil, the NMR mass sensitivity (S/N per mole) increases to a first approximation with 1/d [276;277]. A five fold increase in mass sensitivity, which corresponds to a 25-fold reduction in measurement time compared to a conventional 5 mm probe, was reported for a 1 mm highresolution probe with 2.5 μl active sample volume [278;279]. Other advantages of using low-volume probes are a reduction of approximately a factor 100 in the use of expensive deuterated solvents and a decreased residual solvent signal [278]. Furthermore, microprobes are very well suited for hyphenation with several separation techniques, including HPLC and capillary electrophoresis [280]. In body fluid NMR, the application of microprobes is very valuable when only a minor amount of fluid is available. Its effective use in studying a few microliters of rodent CSF and blood plasma has been demonstrated [278;279]. An additional 190 Future perspectives advantage when studying metabolic disorders in small animal models (e.g. in “knock-out” animals) is that enough body fluid material for microprobe NMR investigations can be obtained without sacrificing the animal. Although microprobes have not yet been applied to study inborn errors of metabolism in man, their enhanced sensitivity and small sample volumes can be promising in this field as well. DIFFUSION EDITED NMR SPECTROSCOPY The presence of macromolecules in blood plasma, serum and CSF can hamper NMR analysis and they are therefore often removed prior to NMR measurement, as described previously. Unfortunately, this can result in loss of metabolites and furthermore inhibit the study of many possible interactions. Diffusion edited spectroscopy applied to intact samples provides an alternative for spectral simplification [281;282]. Under normal conditions, molecular diffusion is not proportional to observed spectral intensities. However, when pulsed field gradients are applied during the experiment, diffusion can cause changes in signal intensities. When compared to a conventional spectrum, fast diffusing molecules such as low molecular weight metabolites will show a much larger intensity difference than slowly diffusing molecules such as lipoproteins. Diffusion edited spectroscopy can therefore effectively suppress small molecule resonances by a factor 100 or more, making it possible to directly study macromolecules [282;283]. A spectrum displaying only low molecular weight metabolites is obtained by subtracting the macromolecular spectrum from a total spectrum recorded without gradients. For a plasma sample, the resulting spectrum is highly similar to a spectrum obtained after plasma deproteinization by ultrafiltration [282]. Since diffusion edited NMR does not require any sample preparation, the experimental time required can be significantly reduced (up to 30 minutes in the case of ultrafiltration), making the use of NMR as a diagnostic tool in clinical chemistry more attractive. The application of diffusion edited spectra to aid resonance assignment in studies of diseases has also been suggested [281;284]. Additionally, a combination of diffusion edited NMR spectroscopy with chemometric methods in toxicological investigations has been reported [285]. In contrast to spin-echo sequences, which can be applied to filter out macromolecule resonances as well, metabolites can be readily quantified with diffusion edited NMR spectroscopy. However, TSP-d4 cannot be used as a concentration reference in this case, since it is partially invisible due to its interaction with proteins. Formate can for instance be used as an alternative. 191 Chapter 5.2 AUTOMATION In complex mixtures like body fluids, each component generates its own NMR spectrum, which possibly overlaps with other compounds in the mixture and makes analysis often difficult. Besides using liquid chromatography coupled to NMR as discussed in paragraph 4.2, this problem can also be solved with automated signal acquisition and special analysis software combined with spectral databases, which can greatly facilitate spectral interpretation and diagnosis of inborn errors of metabolism [286]. Automated NMR measurements require that the apparatus is equipped with a sample changer and that it is capable of self locking and shimming. Nowadays, probes that also perform automatic tuning and matching are becoming available as well. Next, the computer has to start acquisition of the appropriate experiments. In one suggested automation scheme, first a 1D 1H spectrum is recorded and processed, which is then used to detect deviations from normal spectra by principal component analysis incorporating a database with control spectra [286]. Subsequently, a two-dimensional COSY is recorded to identify metabolites that are indicative of the possible inborn errors of metabolism. The success of this approach was demonstrated by correct diagnosis of a few inherited metabolic diseases [286]. A somewhat different set-up used only 1D 1H spectra to directly identify compounds by chemical shift and database comparison [287]. Specialized deconvolution software performs the automatic peak assignment and subsequent metabolite quantification based on peak areas. In total, 1000 urine samples, including 69 samples from patients with 13 different inborn errors of metabolism, were analyzed in less than 2 minutes per sample. All healthy controls were classified as normal, whereas 66 of the pathological samples were correctly diagnosed [287]. A third chemometric method capable of automated analysis is a neural network trained to distinguish 1D 1H spectra of patients with a specific disorder from healthy individuals [59;288]. A neural network is created by supplying data for which the outcome is already known and can include many different diseases. A diagnostic accuracy of almost 100% has been accomplished in the case of the metabolic disorder cystinuria [59]. Automated spectral acquisition and analysis can greatly accelerate spectrum interpretation and facilitate clinical diagnosis by providing clear “yes or no” answers. A second advantage is that it permits NMR measurements to be performed by non-specialists. Therefore, it is likely that further development of these methods can make NMR spectroscopy more attractive as a routine diagnostic tool in hospital laboratories. Although all new developments discussed above show promising results, it is expected that especially automated data acquisition and analysis will provide a large step forward in making NMR spectroscopy a suitable method for standard clinical analysis. It is also likely that a combination of two or more of the described 192 Future perspectives techniques will be employed within several years, incorporating each of their separate advantages. 193 194 Reference list Reference List 1. 2. 3. Schmidt CW. Metabolomics: what's happening downstream of DNA. Environ Health Perspect 2004;112:410-5. The Metabolic and Molecular Bases of Inherited Disease, 8 ed. New York: McGraw-Hill Professional, 2001. Wevers RA, Engelke U, Wendel U, de Jong JG, Gabreëls FJ, Heerschap A. Standardized method for high-resolution 1H-NMR of cerebrospinal fluid. Clin Chem 1995;41:744-51. 4. 5. Wevers RA, Engelke U, Heerschap A. High-resolution 1H-NMR spectroscopy of blood plasma for metabolic studies. Clin Chem 1994;40:1245-50. Engelke UF, Moolenaar SH, Hoenderop SMGC, Morava E, van der Graaf M, Heerschap A, Wever RA. Handbook of 1H-NMR spectroscopy in inborn errors of metabolism: body fluid NMR spectroscopy and in vivo MR spectroscopy, 2 ed. Heilbronn: SPS Verlagsgesellschaft, 2007. 6. Lindon JC, Nicholson JK, Everett JR. NMR spectroscopy of biofluids. Webb GA, editor Annual reports on NMR spectroscopy 7. Nicholson JK, Wilson ID. High Resolution Proton Magnetic Resonance Spectroscopy of Biological Fluids. Progress in Nuclear 8. London: Academic Press 1999;38:1-8. Magnetic Resonance Spectroscopy 1989;21:449-501. Lindon JC, Nicholson JK, Holmes E, Everett JR. Metabonomics: metabolic processes studied by NMR spectroscopy of biofluids. Concepts in Magnetic Resonance 2000;12:289-320. 9. 10. Moolenaar SH, Engelke UF, Wevers RA. Proton nuclear magnetic resonance spectroscopy of body fluids in the field of inborn errors of metabolism. Ann Clin Biochem 2003;40:16-24. Engelke UF, Tangerman A, Willemsen MA, Moskau D, Loss S, Mudd SH, Wevers RA. Dimethyl sulfone in human cerebrospinal fluid and blood plasma confirmed by one-dimensional (1)H and two-dimensional (1)H-(13)C NMR. NMR Biomed 2005;18:331-6. 11. Engelke UF, Liebrand van Sambeek ML, de Jong JG, Leroy JG, Morava E, Smeitink JA, Wevers RA. N-acetylated metabolites in urine: proton nuclear magnetic resonance spectroscopic study on patients with inborn errors of metabolism. Clin Chem 2004;50:58-66. 12. 13. Nicholson JK, Buckingham MJ, Sadler PJ. High resolution 1H n.m.r. studies of vertebrate blood and plasma. Biochem J 1983;211:605-15. Kriat M, Confort Gouny S, Vion Dury J, Sciaky M, Viout P, Cozzone PJ. Quantitation of metabolites in human blood serum by proton magnetic resonance spectroscopy. A comparative study of the use of formate and TSP as concentration standards. NMR Biomed 1992;5:179-84. 14. Connor S, Everett J, Nicholson JK. Spin-echo proton NMR spectroscopy of urine samples. Water suppression via a ureadependent T2 relaxation process. Magn Reson Med 1987;4:461-70. 15. 16. Maillet S, Vion-Dury J, Confort-Gouny S, Nicoli F, Lutz NW, Viout P, Cozzone P. Experimental protocol for clinical analysis of cerebrospinal fluid by high resolution proton magnetic resonance spectroscopy. Brain Res Protocols 1998;3:123-34. Lutz NW, Maillet S, Nicoli F, Viout P, Cozzone PJ. Further assignment of resonances in 1H NMR spectra of cerebrospinal fluid (CSF). FEBS Lett 1998;425:345-51. 17. 18. 19. Sweatman BC, Farrant RD, Holmes E, Ghauri FY, Nicholson JK, Lindon JC. 600 MHz 1H-NMR spectroscopy of human cerebrospinal fluid: effects of sample manipulation and assignment of resonances. J Pharm Biomed Anal 1993;11:651-64. Subramanian A, Gupta A, Saxena S, Gupta A, Kumar R, Nigam A, Kumar R, Mandal SK, Roy R. Proton MR CSF analysis and a new software as predictors for the differentiation of meningitis in children. NMR Biomed (in press) 2005. Bell JD, Brown JC, Sadler PJ, Macleod AF, Sonksen PH, Hughes RD, Williams R. High resolution proton nuclear magnetic resonance studies of human cerebrospinal fluid. Clin Sci (Lond) 1987;72:563-70. 195 Reference list 20. 21. Commodari F, Arnold DL, Sanctuary BC, Shoubridge EA. 1H NMR characterization of normal human cerebrospinal fluid and the detection of methylmalonic acid in a vitamin B12 deficient patient. NMR Biomed 1991;4:192-200. Daykin CA, Foxall PJ, Connor SC, Lindon JC, Nicholson JK. The comparison of plasma deproteinization methods for the detection of low-molecular-weight metabolites by (1)H nuclear magnetic resonance spectroscopy. Anal Biochem 2002;304:220-30. 22. Bell JD, Brown JC, Kubal G, Sadler PJ. NMR-invisible lactate in blood plasma. FEBS Lett 1988;235:81-6. 23. Nicholson JK, Foxall PJ, Spraul M, Farrant RD, Lindon JC. 750 MHz 1H and 1H-13C NMR spectroscopy of human blood plasma. Anal Chem 1995;67:793-811. 24. 25. Lehnert W, Hunkler D. Possibilities of selective screening for inborn errors of metabolism using high-resolution 1H-FT- NMR spectrometry. Eur J Pediatr 1986;145:260-6. Wolf NI, Willemsen MA, Engelke UF, van der Knaap MS, Pouwels PJ, Harting I, Zschocke J, Sistermans EA, Rating D, Wevers RA. Severe hypomyelination associated with increased levels of N-acetylaspartylglutamate in CSF. Neurology 2004;62:1503-8. 26. Sze DY, Jardetzky O. Determination of metabolite and nucleotide concentrations in proliferating lymphocytes by 1H-NMR 27. Holmes E, Foxall PJ, Spraul M, Farrant RD, Nicholson JK, Lindon JC. 750 MHz 1H NMR spectroscopy characterisation of the of acid extracts. Biochim Biophys Acta 1990;1054:181-97. complex metabolic pattern of urine from patients with inborn errors of metabolism: 2-hydroxyglutaric aciduria and maple syrup urine disease. J Pharm Biomed Anal 1997;15:1647-59. 28. Sidelmann UG, Braumann U, Hofmann M, Spraul M, Lindon JC, Nicholson JK, Hansen SH. Directly coupled 800 MHz HPLC- NMR spectroscopy of urine and its applications to the identification of the major phase II metabolites of tolfenamic acid. Anal Chem 1997;69:607-12. 29. Cavanagh J, Fairbrother WJ, Palmer III AG, Skelton NJ. Protein NMR spectroscopy. Principles and practice. Academic Press, 1995. 30. 31. Wijmenga SS, Van Buuren BNM. The use of NMR methods for conformational studies of nucleic acids. Prog Nucl Magn Reson Spec 1998;32:287-387. Van de Ven FJM. Multidimensional NMR in liquids. Basic principles and experimental methods. Wiley-VCH Verlag GmbH, 1995. 32. 33. Iles RA, Jago JR, Williams SR, Chalmers RA. 3-Hydroxy-3-methylglutaryl-CoA lyase deficiency studied using 2- dimensional proton nuclear magnetic resonance spectroscopy. FEBS Lett 1986;203:49-53. Moolenaar SH, van der Knaap MS, Engelke UF, Pouwels PJ, Janssen Zijlstra FS, Verhoeven NM, Jakobs C, Wevers RA. In vivo and in vitro NMR spectroscopy reveal a putative novel inborn error involving polyol metabolism. NMR Biomed 2001;14:167-76. 34. Sewell AC, Murphy HC, Iles RA. Proton nuclear magnetic resonance spectroscopic detection of sialic acid storage disease. Clin Chem 2002;48:357-9. 35. Spraul M, Nicholson JK, Lynch MJ, Lindon JC. Application of the one-dimensional TOCSY pulse sequence in 750 MHz 1H- NMR spectroscopy for assignment of endogenous metabolite resonances in biofluids. J Pharm Biomed Anal 1994;12:6138. 36. Moolenaar SH, Gohlich Ratmann G, Engelke UF, Spraul M, Humpfer E, Dvortsak P, Voit T, Hoffmann GF, Brautigam C, van Kuilenburg AB, van Gennip A, Vreken P, Wevers RA. beta-Ureidopropionase deficiency: a novel inborn error of metabolism discovered using NMR spectroscopy on urine. Magn Reson Med 2001;46:1014-7. 37. Foxall PJ, Parkinson JA, Sadler IH, Lindon JC, Nicholson JK. Analysis of biological fluids using 600 MHz proton NMR spectroscopy: application of homonuclear two-dimensional J-resolved spectroscopy to urine and blood plasma for spectral simplification and assignment. J Pharm Biomed Anal 1993;11:21-31. 38. Williams SR, Iles RA, Chalmers RA. Spin-echo and 2-dimensional 1H nuclear magnetic resonance studies on urinary metabolites from patients with 2-methylacetoacetyl CoA thiolase deficiency. Clin Chim Acta 1986;159:153-61. 196 Reference list 39. 40. Iles RA, Hind AJ, Chalmers RA. Use of proton nuclear magnetic resonance spectroscopy in detection and study of organic acidurias. Clin Chem 1985;31:1795-801. Fan TWM. Metabolite profiling by one- and two-dimensional NMR analysis of complex mixtures. Prog Nucl Magn Reson Spec 1996;28:161-219. 41. Wehrli SL, Berry GT, Palmieri M, Mazur A, Elsas L, Segal S. Urinary galactonate in patients with galactosemia: quantitation 42. Moolenaar SH, Poggi Bach J, Engelke UF, Corstiaensen JM, Heerschap A, de Jong JG, Binzak BA, Vockley J, Wevers RA. by nuclear magnetic resonance spectroscopy. Pediatr Res 1997;42:855-61. Defect in dimethylglycine dehydrogenase, a new inborn error of metabolism: NMR spectroscopy study. Clin Chem 1999;45:459-64. 43. Bal D, Gryff Keller A. 1H and 13C NMR study of 2-hydroxyglutaric acid and its lactone. Magn Reson Chem 2002;40:533-6. 44. Bal D, Gryff Keller A, Gradowska W. Absolute configuration of N-acetylaspartate in urine from patients with Canavan disease. J Inherit Metab Dis 2005;28:607-9. 45. 46. Krawczyk H, Gradowska W. Characterisation of the 1H and 13C NMR spectra of N-acetylaspartylglutamate and its detection in urine from patients with Canavan disease. J Pharm Biomed Anal 2003;31:455-63. Krawczyk H, Gryff Keller A, Gradowska W, Duran M, Pronicka E. 13C NMR spectroscopy: a convenient tool for detection of argininosuccinic aciduria. J Pharm Biomed Anal 2001;26:401-8. 47. Engelke UF, Kremer B, Kluijtmans LA, van der Graaf M, Morava E, Loupatty FJ, Wanders RJ, Moskau D, Loss S, van den Bergh E, Wevers RA. NMR spectroscopic studies on the late onset form of 3-methylglutaconic aciduria type I and other defects in leucine metabolism. NMR Biomed 2006;19:271-8. 48. Holmes E, Foxall PJ, Nicholson JK, Neild GH, Brown SM, Beddell CR, Sweatman BC, Rahr E, Lindon JC, Spraul M. Automatic data reduction and pattern recognition methods for analysis of 1H nuclear magnetic resonance spectra of human urine from normal and pathological states. Anal Biochem 1994;220:284-96. 49. Zuppi C, Messana I, Forni F, Rossi C, Pennacchietti L, Ferrari F, Giardina B. 1H NMR spectra of normal urines: reference ranges of the major metabolites. Clin Chim Acta 1997;265:85-97. 50. Davies SEC, Woolf DA, Chalmers RA, Rafter JEM, Iles RA. Proton nmr studies of betaine excretion in the human neonate: 51. Tang H, Wang Y, Nicholson JK, Lindon JC. Use of relaxation-edited one-dimensional and two dimensional nuclear consequences for choline and methyl group supply. J Nutr Biochem 1992;3:523-30. magnetic resonance spectroscopy to improve detection of small metabolites in blood plasma. Anal Biochem 2004;325:260-72. 52. 53. 54. Petroff OA, Yu RK, Ogino T. High-resolution proton magnetic resonance analysis of human cerebrospinal fluid. J Neurochem 1986;47:1270-6. Sass JO, Schweitzer-Krantz S, Moebus R, Engelke UF, Wevers RA, Omran H. Aminoacylase I deficiency: A "new" inborn error of metabolism? J Inherit Metab Dis 2005;28:50. Sass JO, Mohr V, Olbrich H, Engelke UF, Horvath J, Fliegauf M, Loges NT, Schweitzer-Krantz S, Moebus R, Weiler P, Kispert A, Superti-Furga A, Wevers RA, Omran H. Mutations in ACY1, the Gene Encoding Aminoacylase 1, Cause a Novel Inborn Error of Metabolism. Am J Hum Genet 2006;78:401-9. 55. Van Coster RN, Gerlo EA, Giardina TG, Engelke UF, Smet JE, De Praeter CM, Meersschaut VA, De Meirleir LJ, Seneca SH, Devreese B, Leroy JG, Herga S, Perrier JP, Wevers RA, Lissens W. Aminoacylase I deficiency: a novel inborn error of metabolism. Biochem Biophys Res Commun 2005;338:1322-6. 56. Van Coster RN, Lissens W, Giardina T, Engelke UF, Smet J, Seneca S, De Praeter CM, Herga S, Perrier J, Wevers R, Gerlo E. An Infant with large amounts of several N-acetylated amino acids in the urine has aminoacylase 1 deficiency. J Inherit Metab Dis 2005;28:50. 57. Burns SP, Woolf DA, Leonard JV, Iles RA. Investigation of urea cycle enzyme disorders by 1H-NMR spectroscopy. Clin Chim Acta 1992;209:47-60. 197 Reference list 58. 59. 60. 61. Burns SP, Iles RA. An investigation of argininosuccinic acid anhydrides in argininosuccinic acid lyase deficiency by 1H-NMR spectroscopy. Clin Chim Acta 1993;221:1-13. Bamforth FJ, Dorian V, Vallance H, Wishart DS. Diagnosis of inborn errors of metabolism using 1H NMR spectroscopic analysis of urine. J Inherit Metab Dis 1999;22:297-301. Pontoni G, Rotondo F, Spagnuolo G, Aurino MT, Carteni Farina M, Zappia V, Lama G. Diagnosis and follow-up of cystinuria: use of proton magnetic resonance spectroscopy. Amino Acids 2000;19:469-76. Lehnert W, Stogmann W, Engelke U, Wevers RA, van den Berg GB. Long-term follow up of a new case of hawkinsinuria. Eur J Pediatr 1999;158:578-82. 62. Brown JC, Mills GA, Sadler PJ, Walker V. 1H NMR studies of urine from premature and sick babies. Magn Reson Med 63. Burns SP, Iles RA, Ryalls M, Leonard JV. Methylgenesis from betaine in cystathionine-beta-synthase deficiency. Biochem 1989;11:193-201. Soc Trans 1993;21:455S. 64. Lundberg P, Dudman NP, Kuchel PW, Wilcken DE. 1H NMR determination of urinary betaine in patients with premature vascular disease and mild homocysteinemia. Clin Chem 1995;41:275-83. 65. Yamaguchi S, Koda N, Eto Y, Aoki K. Rapid screening of metabolic disease by proton NMR urinalysis. Lancet 1984;2:284. 66. Moolenaar SH, Engelke UF, Abeling NG, Mandel H, Duran M, Wevers RA. Prolidase deficiency diagnosed by 1H NMR 67. Farrant RD, Walker V, Mills GA, Mellor JM, Langley GJ. Pyridoxal phosphate de-activation by pyrroline-5-carboxylic acid. spectroscopy of urine. J Inherit Metab Dis 2001;24:843-50. Increased risk of vitamin B6 deficiency and seizures in hyperprolinemia type II. J Biol Chem 2001;276:15107-16. 68. De Praeter CM, Gerwig GJ, Bause E, Nuytinck LK, Vliegenthart JF, Breuer W, Kamerling JP, Espeel MF, Martin JJ, De Paepe AM, Chan NW, Dacremont GA, Van Coster RN. A novel disorder caused by defective biosynthesis of N-linked oligosaccharides due to glucosidase I deficiency. Am J Hum Genet 2000;66:1744-56. 69. 70. 71. 72. Onkenhout W, Groener JE, Verhoeven NM, Yin C, Laan LA. L-Arabinosuria: a new defect in human pentose metabolism. Mol Genet Metab 2002;77:80-5. van der Knaap MS, Wevers RA, Struys EA, Verhoeven NM, Pouwels PJ, Engelke UF, Feikema W, Valk J, Jakobs C. Leukoencephalopathy associated with a disturbance in the metabolism of polyols. Ann Neurol 1999;46:925-8. Baumkotter J, Cantz M, Mendla K, Baumann W, Friebolin H, Gehler J, Spranger J. N-Acetylneuraminic acid storage disease. Hum Genet 1985;71:155-9. Bal D, Gradowska W, Gryff Keller A. Determination of the absolute configuration of 2-hydroxyglutaric acid and 5oxoproline in urine samples by high-resolution NMR spectroscopy in the presence of chiral lanthanide complexes. J Pharm Biomed Anal 2002;28:1061-71. 73. Neild GH, Foxall PJ, Lindon JC, Holmes EC, Nicholson JK. Uroscopy in the 21st century: high-field NMR spectroscopy. 74. Mourmans J, Bakkeren J, de Jong J, Wevers R, van Diggelen OP, Suormala T, Baumgartner R, Wendel U. Isolated (biotin- Nephrol Dial Transplant 1997;12:404-17. resistant) 3-methylcrotonyl-CoA carboxylase deficiency: four sibs devoid of pathology. J Inherit Metab Dis 1995;18:6435. 75. Ghauri FY, Parkes HG, Nicholson JK, Wilson ID, Brenton DP. Asymptomatic 5-oxoprolinuria detected by proton magnetic resonance spectroscopy. Clin Chem 1993;39:1341. 76. Yamaguchi S, Koda N, Yamamoto H. Analysis for homogentisic acid by NMR spectrometry, to aid diagnosis of 77. Yamaguchi S, Koda N, Ohashi T. Diagnosis of alkaptonuria by NMR urinalysis: rapid qualitative and quantitative analysis of 198 alkaptonuria. Clin Chem 1989;35:1806-7. homogentisic acid. Tohoku J Exp Med 1986;150:227-8. Reference list 78. 79. 80. 81. Morava E, Kosztolanyi G, Engelke UF, Wevers RA. Reversal of clinical symptoms and radiographic abnormalities with protein restriction and ascorbic acid in alkaptonuria. Ann Clin Biochem 2003;40:108-11. Burlina AP, Kunnecke B., Seelig J, Burlina AB. 1H-NMR spectroscopy of cerebrospinal fluid and urine in Canavan's disease. Proceedings of the 11th Annual Meeting os Society of Magnetic Resonance in Medicine 1992;2:3416. Rafter JE, Chalmers RA, Johnson A, Iles RA. Diagnosis of respiratory chain defects using 1H n.m.r. spectroscopy in vitro. Biochem Soc Trans 1990;18:551-2. Burns SP, Holmes HC, Chalmers RA, Johnson A, Iles RA. Proton NMR spectroscopic analysis of multiple acyl-CoA dehydrogenase deficiency--capacity of the choline oxidation pathway for methylation in vivo. Biochim Biophys Acta 1998;1406:274-82. 82. Yamaguchi S, Koda N, Eto Y, Aoki K. Quick screening and diagnosis of organic acidemia by NMR urinalysis. J Pediatr 1985;106:620-2. 83. Lorek AK, Penrice JM, Cady EB, Leonard JV, Wyatt JS, Iles RA, Burns SP, Reynolds EO. Cerebral energy metabolism in 84. Krawczyk H, Gradowska W. 1H NMR spectra of methylcitric acid in urine. J Inherit Metab Dis 2007;30:263. 85. Kodama S, Sugiura M, Nakao H, Kobayashi K, Miyoshi M, Yoshii K, Komatsu M, Sakurai T. 1H-NMR studies of urine in isovaleric acidaemia. Arch Dis Child Fetal Neonatal Ed 1996;74:F211-F213. propionic acidemia and methylmalonic acidemia. Acta Paediatr Jpn 1991;33:139-45. 86. 87. 88. Iles RA, Chalmers RA, Hind AJ. Methylmalonic aciduria and propionic acidaemia studied by proton nuclear magnetic resonance spectroscopy. Clin Chim Acta 1986;161:173-89. Iles RA, Chalmers RA. Nuclear magnetic resonance spectroscopy in the study of inborn errors of metabolism. Clin Sci (Lond) 1988;74:1-10. Davies SE, Iles RA, Stacey TE, de Sousa C, Chalmers RA. Carnitine therapy and metabolism in the disorders of propionyl- CoA metabolism studied using 1H-NMR spectroscopy. Clin Chim Acta 1991;204:263-77. 89. Oostendorp M, Engelke UF, Willemsen MA, Wevers RA. Diagnosing inborn errors of lipid metabolism with proton nuclear 90. Rafter JE, Chalmers RA, Iles RA. Medium-chain acyl-CoA dehydrogenase deficiency: a 1H-n.m.r. spectroscopic study. 91. Wevers RA, Engelke UF, Moolenaar SH, Brautigam C, de Jong JG, Duran R, de Abreu RA, van Gennip AH. 1H-NMR 92. magnetic resonance spectroscopy. Clin Chem 2006;52:1395-405. Biochem Soc Trans 1990;18:912-3. spectroscopy of body fluids: inborn errors of purine and pyrimidine metabolism. Clin Chem 1999;45:539-48. Assmann B, Gohlich-Ratmann G, Brautigam C, Wagner L, Moolenaar S, Engelke U, Wevers R, Voit T, Hoffmann GF. Presumptive ureidopropionase deficiency as a new defect in pyrimidine catabolism found with in vitro H-NMR spectroscopy. J Inherit Metab Dis 1998;21 (Suppl. 2). 93. Wevers RA, Engelke U, Rotteveel JJ, Heerschap A, de Jong JG, Abeling NG, van Gennip AH, de Abreu RA. 1H NMR spectroscopy of body fluids in patients with inborn errors of purine and pyrimidine metabolism. J Inherit Metab Dis 1997;20:345-50. 94. 95. Schmidt C, Engelke UF, Wevers RA, Anninos A, Feyh P, Hofmann U, Schwab M, Hoffmann GF. Proficiency testing results of different methods of pyrimidine analysis. J Inherit Metab Dis 2004;27:235. Van Hove JL, Damme Lombaerts R, Grunewald S, Peters H, Van Damme B, Fryns JP, Arnout J, Wevers R, Baumgartner ER, Fowler B. Cobalamin disorder Cbl-C presenting with late-onset thrombotic microangiopathy. Am J Med Genet 2002;111:195-201. 96. Binzak BA, Wevers RA, Moolenaar SH, Lee YM, Hwu WL, Poggi Bach J, Engelke UF, Hoard HM, Vockley JG, Vockley J. Cloning of dimethylglycine dehydrogenase and a new human inborn error of metabolism, dimethylglycine dehydrogenase deficiency. Am J Hum Genet 2001;68:839-47. 199 Reference list 97. Tassani M, Valensin G, Hayek J, Engelke UF, Wevers RA. Dimethylglycine dehydrogenase deficiency. J Inherit Metab Dis 98. Schulze A, Hess T, Wevers R, Mayatepek E, Bachert P, Marescau B, Knopp MV, De Deyn PP, Bremer HJ, Rating D. Creatine 2005;28:257. deficiency syndrome caused by guanidinoacetate methyltransferase deficiency: diagnostic tools for a new inborn error of metabolism. J Pediatr 1997;131:626-31. 99. Nicoli F, Vion Dury J, Maloteaux JM, Delwaide C, Confort Gouny S, Sciaky M, Cozzone PJ. CSF and serum metabolic profile of patients with Huntington's chorea: a study by high resolution proton NMR spectroscopy and HPLC. Neurosci Lett 1993;154:47-51. 100. Boltshauser E, Schmitt B, Wevers RA, Engelke U, Burlina AB, Burlina AP. Follow-up of a child with hypoacetylaspartia. 101. Abeling NG, van Gennip AH, Bakker HD, Heerschap A, Engelke U, Wevers RA. Diagnosis of a new case of trimethylaminuria Neuropediatrics 2004;35:255-8. using direct proton NMR spectroscopy of urine. J Inherit Metab Dis 1995;18:182-4. 102. 103. Maschke S, Wahl A, Azaroual N, Boulet O, Crunelle V, Imbenotte M, Foulard M, Vermeersch G, Lhermitte M. 1H-NMR analysis of trimethylamine in urine for the diagnosis of fish-odour syndrome. Clin Chim Acta 1997;263:139-46. Mazon Ramos A, Gil Setas A, Berrade Zubiri S, Bandres Echeverri T, Wevers R, Engelke U, Zschocke J. Trimetilaminuria o sindrome del olor a pescado. Nueva mutacion genica y primer caso documentado en Espana. [Primary trimethylaminuria or fish odor syndrome. A novel mutation in the first documented case in Spain]. Med Clin (Barc) 2003;120:219-21. 104. Podadera P, Areas JA, Lanfer Marquez UM. Diagnosis of suspected trimethylaminuria by NMR spectroscopy. Clin Chim Acta 2005;351:149-54. 105. Murphy HC, Dolphin CT, Janmohamed A, Holmes HC, Michelakakis H, Shephard EA, Chalmers RA, Phillips IR, Iles RA. A novel mutation in the flavin-containing monooxygenase 3 gene, FM03, that causes fish-odour syndrome: activity of the mutant enzyme assessed by proton NMR spectroscopy. Pharmacogenetics 2000;10:439-51. 106. 107. 108. 109. Matsushita K, Kato K, Ohsaka A, Kanazawa M, Aizawa K. A simple and rapid method for detecting trimethylamine in human urine by proton NMR. Physiol Chem Phys Med NMR 1989;21:3-4. Eugene M. Diagnostic du "fish odour syndrome" par spectrometrie RMN proton des urines. [Diagnosis of "fish odor syndrome" by urine nuclear magnetic resonance proton spectrometry]. Ann Dermatol Venereol 1998;125:210-2. Nicholson JK, Gartland KP. 1H NMR studies on protein binding of histidine, tyrosine and phenylalanine in blood plasma. NMR Biomed 1989;2:77-82. Hochuli M, Wuthrich K, Steinmann B. Two-dimensional NMR spectroscopy of urinary glycosaminoglycans from patients with different mucopolysaccharidoses. NMR Biomed 2003;16:224-36. 110. Blau N, Duran M, Blaskovics ME, Gibson KM. Physician's guide to the laboratory diagnostics of metabolic diseases, 2 ed. 111. al Waiz M, Ayesh R, Mitchell SC, Idle JR, Smith RL. Trimethylaminuria: the detection of carriers using a trimethylamine load Berlin Heidelberg: Springer-Verlag, 2003. test. J Inherit Metab Dis 1989;12:80-5. 112. 113. Marks R, Greaves MW, Prottey C, Hartop PJ. Trimethylaminuria: the use of choline as an aid to diagnosis. Br J Dermatol 1977;96:399-402. Schulze A, Ebinger F, Rating D, Mayatepek E. Improving treatment of guanidinoacetate methyltransferase deficiency: reduction of guanidinoacetic acid in body fluids by arginine restriction and ornithine supplementation. Mol Genet Metab 2001;74:413-9. 114. Vreken P, van Kuilenburg AB, Hamajima N, Meinsma R, van Lenthe H, Gohlich Ratmann G, Assmann BE, Wevers RA, van Gennip AH. cDNA cloning, genomic structure and chromosomal localization of the human BUP-1 gene encoding betaureidopropionase. Biochim Biophys Acta 1999;1447:251-7. 115. 200 van Kuilenburg AB, Meinsma R, Beke E, Assmann B, Ribes A, Lorente I, Busch R, Mayatepek E, Abeling NG, van Cruchten A, Stroomer AE, van Lenthe H, Zoetekouw L, Kulik W, Hoffmann GF, Voit T, Wevers RA, Rutsch F, van Gennip AH. beta- Reference list Ureidopropionase deficiency: an inborn error of pyrimidine degradation associated with neurological abnormalities. Hum Mol Genet 2004;13:2793-801. 116. Huck JH, Verhoeven NM, Struys EA, Salomons GS, Jakobs C, van der Knaap MS. Ribose-5-phosphate isomerase deficiency: new inborn error in the pentose phosphate pathway associated with a slowly progressive leukoencephalopathy. Am J Hum Genet 2004;74:745-51. 117. Meshitsuka S, Koeda T, Muro H. Direct observation of 3-keto-valproate in urine by 2D-NMR spectroscopy. Clin Chim Acta 118. Azaroual N, Imbenotte M, Cartigny B, Leclerc F, Vallee L, Lhermitte M, Vermeersch G. Valproic acid intoxication identified 119. Komoroski EM, Komoroski RA, Valentine JL, Pearce JM, Kearns GL. The use of nuclear magnetic resonance spectroscopy in 120. 2003;334:145-51. by 1H and 1H-(13)C correlated NMR spectroscopy of urine samples. MAGMA 2000;10:177-82. the detection of drug intoxication. J Anal Toxicol 2000;24:180-7. Spraul M, Hofmann M, Lindon JC, Farrant RD, Seddon MJ, Nicholson JK, Wilson ID. Evaluation of liquid chromatography coupled with high-field 1H NMR spectroscopy for drug metabolite detection and characterization: the identification of paracetamol metabolites in urine and bile. NMR Biomed 1994;7:295-303. 121. Godejohann M, Tseng LH, Braumann U, Fuchser J, Spraul M. Characterization of a paracetamol metabolite using on-line LC-SPE-NMR-MS and a cryogenic NMR probe. J Chromatogr A 2004;1058:191-6. 122. Connor SC, Everett JR, Jennings KR, Nicholson JK, Woodnutt G. High resolution 1H NMR spectroscopic studies of the 123. Jaroszewski JW, Berenstein D, Slok FA, Simonsen PE, Agger MK. Determination of diethylcarbamazine, an antifilarial drug, 124. metabolism and excretion of ampicillin in rats and amoxycillin in rats and man. J Pharm Pharmacol 1994;46:128-34. in human urine by 1H-NMR spectroscopy. J Pharm Biomed Anal 1996;14:543-9. Strandholm JJ, Buist NR, Kennaway NG, Curtis HT. Excretion of -N-acetylcitrulline in citrullinaemia. Biochim Biophys Acta 1971;244:214-6. 125. Jellum E, Horn L, Thoresen O, Kvittingen EA, Stokke O. Urinary excretion of N-acetyl amino acids in patients with some 126. Burlina AP, Ferrari V, Divry P, Gradowska W, Jakobs C, Bennett MJ, Sewell AC, Dionisi Vici C, Burlina AB. N- 127. van-Pelt J, Dorland L, Duran M, Hokke CH, Kamerling JP, Vliegenthart JF. Sialyl-alpha 2-6-mannosyl-beta 1-4-N- inborn errors of amino acid metabolism. Scand J Clin Lab Invest Suppl 1986;184:21-6. acetylaspartylglutamate in Canavan disease: an adverse effector? Eur J Pediatr 1999;158:406-9. acetylglucosamine, a novel compound occurring in urine of patients with beta-mannosidosis. J Biol Chem 1990;265:19685-9. 128. van-Pelt J, Hard K, Kamerling JP, Vliegenthart JF, Reuser AJ, Galjaard H. Isolation and structural characterization of twentyone sialyloligosaccharides from galactosialidosis urine. An intact N,N'-diacetylchitobiose unit at the reducing end of a diantennary structure. Biol Chem Hoppe Seyler 1989;370:191-203. 129. van-Pelt J, Kamerling JP, Bakker HD, Vliegenthart JF. A comparative study of sialyloligosaccharides isolated from sialidosis and galactosialidosis urine. J Inherit Metab Dis 1991;14:730-40. 130. 131. Parvy PR, Bardet JI, Rabier DM, Kamoun PP. Age-related reference values for free amino acids in first morning urine specimens. Clin Chem 1988;34:2092-5. Romppanen J, Mononen I. Age-related reference values for urinary excretion of sialic acid and deoxysialic acid: application to diagnosis of storage disorders of free sialic acid. Clin Chem 1995;41:544-7. 132. Hard K, Mekking A, Kamerling JP, Dacremont GA, Vliegenthart JF. Different oligosaccharides accumulate in the brain and urine of a cat with alpha-mannosidosis: structure determination of five brain-derived and seventeen urinary oligosaccharides. Glycoconj J 1991;8:17-28. 133. Michalski JC, Wieruszeski JM, Alonso C, Cache P, Montreuil J, Strecker G. Characterization and 400-MHz 1H-NMR analysis of urinary fucosyl glycoasparagines in fucosidosis. Eur J Biochem 1991;201:439-58. 201 Reference list 134. 135. 136. Pertti Aula, Anu Jalanko, Leena Peltonen. Aspartylglucosaminuria. In: Scriver, Beadet, Valle, Sly, eds. The Metabolic & Molecular Bases of Inherited Diseases. 2001:3535-50. Biocarb. BioCarb Chemicals; Catalog with H-NMR spectra. Lund (Sweden): 1990. Strecker G, Herlant-Peers MC, Fournet B, Montreul J. Structure of seven oligosaccharides excreted in the urine of a patient with Sandhoff's disease (GM2 gangliosidosis-variant O). Eur J Biochem 1977;81:165-71. 137. Dorland L, Haverkamp J, Viliegenthart JF, Strecker G, Michalski JC, Fournet B, Spik G, Montreuil J. 360-MHz 1H nuclearmagnetic-resonance spectroscopy of sialyl-oligosaccharides from patients with sialidosis (mucolipidosis I and II). Eur J Biochem 1978;87:323-9. 138. Jakobs C, Verhoeven NM, van-der-Knaap MS. Various Organic Acidurias. In: Blau N, Duran M, Blaskovics ME, Gibson KM, 139. Ruitenbeek W, Kobayashi K, Iijima M, Smeitink JA, Engelke UF, de Abreu RA, Kwast HT, Saheki T, Boelen CA, de Jong JG, eds. Physician's Guide to the Laboratory Diagnosis of Metabolic Diseases. 2003:225-6. Wevers RA. Moderate citrullinaemia without hyperammonaemia in a child with mutated and deficient argininosuccinate synthetase. Ann Clin Biochem 2003;40:102-7. 140. Blau N, Duran M, Blaskovics ME, Gibson KM. Physician's Guide to the Laboratory Diagnosis of Metabolic Diseases, 2 ed. Berlin Heidelberg NewYork: Springer-Verlag, 2002:38-39pp. 141. van der Knaap MS, Valk J. Magnetic resonance of myelin, myelination, and myelin disorders., 2 ed. Berlin: Springer, 1995. 142. Barkovich AJ. Concepts of myelin and myelination in neuroradiology. AJNR Am J Neuroradiol 2000;21:1099-109. 143. van der Knaap MS, Breiter SN, Naidu S, Hart AA, Valk J. Defining and categorizing leukoencephalopathies of unknown origin: MR imaging approach. Radiology 1999;213:121-33. 144. 145. Beaulieu C, D'Arceuil H, Hedehus M, de Crespigny A, Kastrup A, Moseley ME. Diffusion-weighted magnetic resonance imaging: theory and potential applications to child neurology. Semin Pediatr Neurol 1999;6:87-100. Stockler S, Holzbach U, Hanefeld F, Marquardt I, Helms G, Requart M, Hanicke W, Frahm J. Creatine deficiency in the brain: a new, treatable inborn error of metabolism. Pediatr Res 1994;36:409-13. 146. 147. Martin E, Capone A, Schneider J, Hennig J, Thiel T. Absence of N-acetylaspartate in the human brain: impact on neurospectroscopy? Ann Neurol 2001;49:518-21. Hudson LD. Pelizaeus-Merzbacher disease and the allelic disorder X-linked spastic paraplegia type II. In: Scriver CR, Beaudet AL, Sly WS, Valle D, eds. The metabolic and molecular bases of inherited disease. New York: 2001:5789-98. 148. Hudson LD, Puckett C, Berndt J, Chan J, Gencic S. Mutation of the proteolipid protein gene PLP in a human X chromosome- linked myelin disorder. Proc Natl Acad Sci U S A 1989;86:8128-31. 149. Provencher SW. Estimation of metabolite concentrations from localized in vivo proton NMR spectra. Magn Reson Med 150. Pouwels PJ, Frahm J. Regional metabolite concentrations in human brain as determined by quantitative localized proton 151. Pouwels PJ, Brockmann K, Kruse B, Wilken B, Wick M, Hanefeld F, Frahm J. Regional age dependence of human brain 152. 153. 154. 1993;30:672-9. MRS. Magn Reson Med 1998;39:53-60. metabolites from infancy to adulthood as detected by quantitative localized proton MRS. Pediatr Res 1999;46:474-85. Sistermans EA, de Coo RF, De Wijs IJ, Van Oost BA. Duplication of the proteolipid protein gene is the major cause of Pelizaeus-Merzbacher disease. Neurology 1998;50:1749-54. Faull KF, Rafie R, Pascoe N, Marsh L, Pfefferbaum A. N-acetylaspartic acid (NAA) and N-acetylaspartylglutamic acid (NAAG) in human ventricular, subarachnoid, and lumbar cerebrospinal fluid. Neurochem Res 1999;24:1249-61. van der Knaap MS, Valk J. The reflection of histology in MR imaging of Pelizaeus-Merzbacher disease. AJNR Am J Neuroradiol 1989;10:99-103. 202 Reference list 155. Plecko B, Stockler Ipsiroglu S, Gruber S, Mlynarik V, Moser E, Simbrunner J, Ebner F, Bernert G, Harrer G, Gal A, Prayer D. Degree of hypomyelination and magnetic resonance spectroscopy findings in patients with Pelizaeus Merzbacher phenotype. Neuropediatrics 2003;34:127-36. 156. Nezu A, Kimura S, Takeshita S, Osaka H, Kimura K, Inoue K. An MRI and MRS study of Pelizaeus-Merzbacher disease. Pediatr Neurol 1998;18:334-7. 157. Cassidy SB, Sheehan NC, Farrell DF, Grunnet M, Holmes GL, Zimmerman AW. Connatal Pelizaeus-Merzbacher disease: an 158. Begleiter ML, Harris DJ. Autosomal recessive form of connatal Pelizaeus-Merzbacher disease. Am J Med Genet 159. autosomal recessive form. Pediatr Neurol 1987;3:300-5. 1989;33:311-3. Sistermans EA, De Wijs IJ, de Coo RF, Smit LM, Menko FH, Van Oost BA. A (G-to-A) mutation in the initiation codon of the proteolipid protein gene causing a relatively mild form of Pelizaeus-Merzbacher disease in a Dutch family. Hum Genet 1996;97:337-9. 160. Inoue K, Tanaka H, Scaglia F, Araki A, Shaffer LG, Lupski JR. Compensating for central nervous system dysmyelination: 161. Hodes ME, DeMyer WE, Pratt VM, Edwards MK, Dlouhy SR. Girl with signs of Pelizaeus-Merzbacher disease heterozygous females with a proteolipid protein gene duplication and sustained clinical improvement. Ann Neurol 2001;50:747-54. for a mutation in exon 2 of the proteolipid protein gene. Am J Med Genet 1995;55:397-401. 162. Raskind WH, Williams CA, Hudson LD, Bird TD. Complete deletion of the proteolipid protein gene (PLP) in a family with X- 163. Nance MA, Boyadjiev S, Pratt VM, Taylor S, Hodes ME, Dlouhy SR. Adult-onset neurodegenerative disorder due to linked Pelizaeus-Merzbacher disease. Am J Hum Genet 1991;49:1355-60. proteolipid protein gene mutation in the mother of a man with Pelizaeus-Merzbacher disease. Neurology 1996;47:13335. 164. Sivakumar K, Sambuughin N, Selenge B, Nagle JW, Baasanjav D, Hudson LD, Goldfarb LG. Novel exon 3B proteolipid protein gene mutation causing late-onset spastic paraplegia type 2 with variable penetrance in female family members. Ann Neurol 1999;45:680-3. 165. Pouwels PJ, Hanefeld F, Frahm J. Proton MRS in Pelizaeus-Merzbacher disease. Neuropediatrics 1997;28:355-6. 166. Garbern JY, Yool DA, Moore GJ, Wilds IB, Faulk MW, Klugmann M, Nave KA, Sistermans EA, van der Knaap MS, Bird TD, Shy ME, Kamholz JA, Griffiths IR. Patients lacking the major CNS myelin protein, proteolipid protein 1, develop lengthdependent axonal degeneration in the absence of demyelination and inflammation. Brain 2002;125:551-61. 167. Takanashi J, Inoue K, Tomita M, Kurihara A, Morita F, Ikehira H, Tanada S, Yoshitome E, Kohno Y. Brain N-acetylaspartate is elevated in Pelizaeus-Merzbacher disease with PLP1 duplication. Neurology 2002;58:237-41. 168. 169. Bonavita S, Schiffmann R, Moore DF, Frei K, Choi B, Patronas MD N, Virta A, Boespflug Tanguy O, Tedeschi G. Evidence for neuroaxonal injury in patients with proteolipid protein gene mutations. Neurology 2001;56:785-8. Pizzini F, Fatemi AS, Barker PB, Nagae Poetscher LM, Horska A, Zimmerman AW, Moser HW, Bibat G, Naidu S. Proton MR spectroscopic imaging in Pelizaeus-Merzbacher disease. AJNR Am J Neuroradiol 2003;24:1683-9. 170. Burlina AP, Corazza A, Ferrari V, Erhard P, Kunnecke B, Seelig J, Burlina AB. Detection of increased urinary N- 171. Burlina AP, Ferrari V, Burlina AB, Bertini E, Boespflug-Tanguy O. Increased levels of N-acetylaspartylglutamate in CSF and acetylaspartylglutamate in Canavan disease. Eur J Pediatr 1994;153:538-9. urine of patients with Pelizaeus-Merzbacher disease. J Inhert Metab Dis 2000;23(suppl 1):213. 172. 173. Neale JH, Bzdega T, Wroblewska B. N-Acetylaspartylglutamate: the most abundant peptide neurotransmitter in the mammalian central nervous system. J Neurochem 2000;75:443-52. Baslow MH. Functions of N-acetyl-L-aspartate and N-acetyl-L-aspartylglutamate in the vertebrate brain: role in glial cell- specific signaling. J Neurochem 2000;75:453-9. 203 Reference list 174. 175. Luthi Carter R, Barczak AK, Speno H, Coyle JT. Hydrolysis of the neuropeptide N-acetylaspartylglutamate (NAAG) by cloned human glutamate carboxypeptidase II. Brain Res 1998;795:341-8. Halsted CH, Ling EH, Luthi Carter R, Villanueva JA, Gardner JM, Coyle JT. Folylpoly-gamma-glutamate carboxypeptidase from pig jejunum. Molecular characterization and relation to glutamate carboxypeptidase II. J Biol Chem 1998;273:2041724. 176. Bacich DJ, Ramadan E, O'Keefe DS, Bukhari N, Wegorzewska I, Ojeifo O, Olszewski R, Wrenn CC, Bzdega T, Wroblewska B, Heston WD, Neale JH. Deletion of the glutamate carboxypeptidase II gene in mice reveals a second enzyme activity that hydrolyzes N-acetylaspartylglutamate. J Neurochem 2002;83:20-9. 177. Gingras R, Richard C, El Alfy M, Morales CR, Potier M, Pshezhetsky AV. Purification, cDNA cloning, and expression of a new human blood plasma glutamate carboxypeptidase homologous to N-acetyl-aspartyl-alpha-glutamate carboxypeptidase/prostate-specific membrane antigen. J Biol Chem 1999;274:11742-50. 178. 179. 180. 181. Anders MW, Dekant W. Aminoacylases. Adv Pharmacol 1994;27:431-48. Editors: Charles R.Scriver, William S.Sly. The Metabolic and Molecular Bases of Inherited Disease, 8 ed. New York: McGraw- Hill Professional, 2001. J.O.Sass, et al. NEUROLOGICAL FINDINGS IN AMINOACYLASE 1 DEFICIENCY (Submitted to Neurology). Neurology 2007. Winitz M, Bloch-Frankenthal L, Izumiya N, Birnbaum S, Baker C, Greenstein J. Studies on Diastereoisomeric alfa-Amino Acids and Corresponding alfa-Hydroxy Acids. J Am Chem Soc 1956;78:2423-30. 182. Gerlo E, Van Coster RN, Lissens W, Winckelmans G, De Meirleir LJ, Wevers R. Gas chromatographic-mass spectrometric 183. Engelke UF, Kremer B, Kluijtmans LA, van der Graaf M, Morava E, Loupatty FJ, Wanders RJ, Moskau D, Loss S, van den analysis of N-acetylated amino acids: the first case of aminoacylase I deficiency. Anal Chim Acta 2006;571:191-9. Bergh E, Wevers RA. NMR spectroscopic studies on the late onset form of 3-methylglutaconic aciduria type I and other defects in leucine metabolism. NMR Biomed 2006;19:271-8. 184. Lehnert W. Long-term results of selective screening for inborn errors of metabolism. Eur J Pediatr 1994;153:S9-13. 185. Sugahara K, Zhang J, Kodama H. Liquid chromatographic-mass spectrometric analysis of N-acetylamino acids in human 186. urine. J Chromatogr B Biomed Appl 1994;657:15-21. Polevoda B, Sherman F. N-terminal acetyltransferases and sequence requirements for N-terminal acetylation of eukaryotic proteins. J Mol Biol 2003;325:595-622. 187. Jakobs C, ten Brink HJ, Langelaar SA, Zee T, Stellaard F, Macek M, Srsnova K, Srsen S, Kleijer WJ. Stable isotope dilution analysis of N-acetylaspartic acid in CSF, blood, urine and amniotic fluid: accurate postnatal diagnosis and the potential for prenatal diagnosis of Canavan disease. J Inherit Metab Dis 1991;14:653-60. 188. 189. Chamberlin ME, Ubagai T, Mudd SH, Wilson WG, Leonard JV, Chou JY. Demyelination of the brain is associated with methionine adenosyltransferase I/III deficiency. J Clin Invest 1996;98:1021-7. Kotb M, Mudd SH, Mato JM, Geller AM, Kredich NM, Chou JY, Cantoni GL. Consensus nomenclature for the mammalian methionine adenosyltransferase genes and gene products. Trends Genet 1997;13:51-2. 190. 191. Blom HJ, Boers GH, van den Elzen JP, Gahl WA, Tangerman A. Transamination of methionine in humans. Clin Sci (Lond) 1989;76:43-9. Mudd SH, Levy HL, Tangerman A, Boujet C, Buist N, Davidson Mundt A, Hudgins L, Oyanagi K, Nagao M, Wilson WG. Isolated persistent hypermethioninemia. Am J Hum Genet 1995;57:882-92. 192. Williams RT. "Detoxication Mechanism,", 2 ed. New York: 1959:39pp. 193. Williams KI, Burstein SH, Layne DS. Dimethyl sulfone: isolation from human urine. Arch Biochem Biophys 1966;113:251-2. 194. Jacob SW, Herschler R. Dimethyl sulfoxide after twenty years. Ann N Y Acad Sci 1983;411:xiii-xvii. 204 Reference list 195. Cecil KM, Lin A, Ross BD, Egelhoff JC. Methylsulfonylmethane observed by in vivo proton magnetic resonance spectroscopy in a 5-year-old child with developmental disorder: effects of dietary supplementation. J Comput Assist Tomogr 2002;26:818-20. 196. Lin A, Nguy CH, Shic F, Ross BD. Accumulation of methylsulfonylmethane in the human brain: identification by multinuclear magnetic resonance spectroscopy. Toxicol Lett 2001;123:169-77. 197. Rose SE, Chalk JB, Galloway GJ, Doddrell DM. Detection of dimethyl sulfone in the human brain by in vivo proton magnetic 198. Williams KI, Burstein SH, Layne DS. Dimethyl sulfone: isolation from cows' milk. Proc Soc Exp Biol Med 1966;122:865-6. 199. resonance spectroscopy. Magn Reson Imaging 2000;18:95-8. Silva-Ferreira AC, Rodrigues P, Hogg T, Guedes-De-Pinho P. Influence of some technological parameters on the formation of dimethyl sulfide, 2-mercaptoethanol, methionol, and dimethyl sulfone in port wines. J Agric Food Chem 2003;51:72732. 200. Williams KI, Burstein SH, Layne DS. Metabolism of dimethyl sulfide, dimethyl sulfoxide, and dimethyl sulfone in the rabbit. 201. Scarlata CJ, Ebeler SE. Headspace solid-phase microextraction for the analysis of dimethyl sulfide in beer. J Agric Food 202. 203. Arch Biochem Biophys 1966;117:84-7. Chem 1999;47:2505-8. Ren YL, Desmarchelier JM, Williams P, Delves R. Natural levels of dimethyl sulfide in rough rice and its products. J Agric Food Chem 2001;49:705-9. Kotseridis Y, Baumes R. Identification of impact odorants in Bordeaux red grape juice, in the commercial yeast used for its fermentation, and in the produced wine. J Agric Food Chem 2000;48:400-6. 204. 205. Weisiger RA, Pinkus LM, Jakoby WB. Thiol S-methyltransferase: suggested role in detoxication of intestinal hydrogen sulfide. Biochem Pharmacol 1980;29:2885-7. Furne J, Springfield J, Koenig T, DeMaster E, Levitt MD. Oxidation of hydrogen sulfide and methanethiol to thiosulfate by rat tissues: a specialized function of the colonic mucosa. Biochem Pharmacol 2001;62:255-9. 206. Blom HJ, Tangerman A. Methanethiol metabolism in whole blood. J Lab Clin Med 1988;111:606-10. 207. Layman DL, Jacob SW. The absorption, metabolism and excretion of dimethyl sulfoxide by rhesus monkeys. Life Sci 208. Paulin HJ, Murphy JB, Larson RE. Determination of dimethyl sulfoxide in plasma and cerebrospinal fluid by gas--liquid 209. 1985;37:2431-7. chromatography. Anal Chem 1966;38:651-2. Pouwels PJ, Frahm J. Regional metabolite concentrations in human brain as determined by quantitative localized proton MRS. Magn Reson Med 1998;39:53-60. 210. 211. 212. 213. Methylsulfonylmethane (MSM). Monograph. Altern Med Rev 2003;8:438-41. Fein G, Meyerhoff DJ. Ethanol in human brain by magnetic resonance spectroscopy: correlation with blood and breath levels, relaxation, and magnetization transfer. Alcohol Clin Exp Res 2000;24:1227-35. Kingsley PB, Shah TC, Woldenberg R. Identification of diffuse and focal brain lesions by clinical magnetic resonance spectroscopy. NMR Biomed 2006;19:435-62. Brockmann K, Bjornstad A, Dechent P, Korenke CG, Smeitink J, Trijbels JM, Athanassopoulos S, Villagran R, Skjeldal OH, Wilichowski E, Frahm J, Hanefeld F. Succinate in dystrophic white matter: a proton magnetic resonance spectroscopy finding characteristic for complex II deficiency. Ann Neurol 2002;52:38-46. 214. 215. Varho T, Komu M, Sonninen P, Holopainen I, Nyman S, Manner T, Sillanpaa M, Aula P, Lundbom N. A new metabolite contributing to N-acetyl signal in 1H MRS of the brain in Salla disease. Neurology 1999;52:1668-72. Grodd W, Krageloh Mann I, Petersen D, Trefz FK, Harzer K. In vivo assessment of N-acetylaspartate in brain in spongy degeneration (Canavan's disease) by proton spectroscopy. Lancet 1990;336:437-8. 205 Reference list 216. Willemsen MA, van der Graaf M, van der Knaap MS, Heerschap A, Van Domburg PH, Gabreels FJ, Rotteveel JJ. MR imaging and proton MR spectroscopic studies in Sjogren-Larsson syndrome: characterization of the leukoencephalopathy. AJNR Am J Neuroradiol 2004;25:649-57. 217. Linnankivi T, Lundbom N, Autti T, Hakkinen AM, Koillinen H, Kuusi T, Lonnqvist T, Sainio K, Valanne L, Aarimaa T, Pihko H. Five new cases of a recently described leukoencephalopathy with high brain lactate. Neurology 2004;63:688-92. 218. van der Knaap MS, van der Voorn P, Barkhof F, Van Coster R, Krageloh Mann I, Feigenbaum A, Blaser S, Vles JS, Rieckmann P, Pouwels PJ. A new leukoencephalopathy with brainstem and spinal cord involvement and high lactate. Ann Neurol 2003;53:252-8. 219. Brooks DJ, Lammertsma AA, Beaney RP, Leenders KL, Buckingham PD, Marshall J, Jones T. Measurement of regional cerebral pH in human subjects using continuous inhalation of 11CO2 and positron emission tomography. J Cereb Blood Flow Metab 1984;4:458-65. 220. 221. 222. 223. Kelley RI. Quantification of 3-methylglutaconic acid in urine, plasma, and amniotic fluid by isotope-dilution gas chromatography/mass spectrometry. Clin Chim Acta 1993;220:157-64. Illsinger S, Lucke T, Zschocke J, Gibson KM, Das AM. 3-methylglutaconic aciduria type I in a boy with fever-associated seizures. Pediatr Neurol 2004;30:213-5. Duran M, Baumgartner ER, Suormala TM, Bruinvis L, Dorland L, Smeitink JA, Poll The BT. Cerebrospinal fluid organic acids in biotinidase deficiency. J Inherit Metab Dis 1993;16:513-6. Yudkoff M, Daikhin Y, Grunstein L, Nissim I, Stern J, Pleasure D, Nissim I. Astrocyte leucine metabolism: significance of branched-chain amino acid transamination. J Neurochem 1996;66:378-85. 224. 225. Bixel M, Shimomura Y, Hutson S, Hamprecht B. Distribution of key enzymes of branched-chain amino acid metabolism in glial and neuronal cells in culture. J Histochem Cytochem 2001;49:407-18. Bixel MG, Hamprecht B. Immunocytochemical localization of beta-methylcrotonyl-CoA carboxylase in astroglial cells and neurons in culture. J Neurochem 2000;74:1059-67. 226. 227. Lynen F, Knappe J, Lorch E, Juetting G, Ringelmann E, Lachance J. On the biochemical function of biotin. II. Purification and mode of action of beta-methyl-crotonyl-carboxylase. [Article in German]. Biochem Z 1961;335:123-67. Sweetman L, Williams J. Branched Chain Organic Acidurias. In: Scriver C, Beaudet A, Sly W, Valle D, eds. The Metabolic and Molecular Basis of Inherited Diseases. 2001:2125-63. 228. Ly TB, Peters V, Gibson KM, Liesert M, Buckel W, Wilcken B, Carpenter K, Ensenauer R, Hoffmann GF, Mack M, Zschocke J. 229. IJlst L, Loupatty FJ, Ruiter JP, Duran M, Lehnert W, Wanders RJ. 3-Methylglutaconic aciduria type I is caused by mutations in 230. Loupatty FJ, Ruiter JP, IJlst L, Duran M, Wanders RJ. Direct nonisotopic assay of 3-methylglutaconyl-CoA hydratase in Mutations in the AUH gene cause 3-methylglutaconic aciduria type I. Hum Mutat 2003;21:401-7. AUH. Am J Hum Genet 2002;71:1463-6. cultured human skin fibroblasts to specifically identify patients with 3-methylglutaconic aciduria type I. Clin Chem 2004;50:1447-50. 231. 232. 233. 234. Arbelaez A, Castillo M, Stone J. MRI in 3-methylglutaconic aciduria type 1. Neuroradiology 1999;41:941-2. Shoji Y, Takahashi T, Sawaishi Y, Ishida A, Matsumori M, Shoji Y, Enoki M, Watanabe H, Takada G. 3-Methylglutaconic aciduria type I: clinical heterogeneity as a neurometabolic disease. J Inherit Metab Dis 1999;22:1-8. Hou JW, Wang TR. 3-Methylglutaconic aciduria presenting as Reye syndrome in a Chinese boy. J Inherit Metab Dis 1995;18:645-6. Jooste S, Erasmus E, Mienie LJ, de Wet WJ, Gibson KM. The detection of 3-methylglutarylcarnitine and a new dicarboxylic conjugate, 3-methylglutaconylcarnitine, in 3-methylglutaconic aciduria. Clin Chim Acta 1994;230:1-8. 235. 206 Narisawa K, Gibson KM, Sweetman L, Nyhan WL, Duran M, Wadman SK. Deficiency of 3-methylglutaconyl-coenzyme A hydratase in two siblings with 3-methylglutaconic aciduria. J Clin Invest 1986;77:1148-52. Reference list 236. Duran M, Beemer FA, Tibosch AS, Bruinvis L, Ketting D, Wadman SK. Inherited 3-methylglutaconic aciduria in two 237. Gibson KM, Lee CF, Wappner RS. 3-Methylglutaconyl-coenzyme-A hydratase deficiency: a new case. J Inherit Metab Dis 238. Gibson KM, Wappner RS, Jooste S, Erasmus E, Mienie LJ, Gerlo E, Desprechins B, De Meirleir L. Variable clinical presentation 239. brothers--another defect of leucine metabolism. J Pediatr 1982;101:551-4. 1992;15:363-6. in three patients with 3-methylglutaconyl-coenzyme A hydratase deficiency. J Inherit Metab Dis 1998;21:631-8. Ensenauer R, Muller CB, Schwab KO, Gibson KM, Brandis M, Lehnert W. 3-Methylglutaconyl-CoA hydratase deficiency: a new patient with speech retardation as the leading sign. J Inherit Metab Dis 2000;23:341-4. 240. Anikster Y, Kleta R, Shaag A, Gahl WA, Elpeleg O. Type III 3-methylglutaconic aciduria (optic atrophy plus syndrome, or Costeff optic atrophy syndrome): identification of the OPA3 gene and its founder mutation in Iraqi Jews. Am J Hum Genet 2001;69:1218-24. 241. Holden HM, Benning MM, Haller T, Gerlt JA. The crotonase superfamily: divergently related enzymes that catalyze different 242. Bell AF, Wu J, Feng Y, Tonge PJ. Involvement of glycine 141 in substrate activation by enoyl-CoA hydratase. Biochemistry 243. reactions involving acyl coenzyme a thioesters. Acc Chem Res 2001;34:145-57. 2001;40:1725-33. Kurimoto K, Fukai S, Nureki O, Muto Y, Yokoyama S. Crystal structure of human AUH protein, a single-stranded RNA binding homolog of enoyl-CoA hydratase. Structure (Camb) 2001;9:1253-63. 244. Ala-Korpela M. 1H NMR spectroscopy of human blood plasma. Prog Nucl Magn Reson 1995;27:475-554. 245. Wilson WK, Sumpter RM, Warren JJ, Rogers PS, Ruan B, Schroepfer GJJ. Analysis of unsaturated C27 sterols by nuclear 246. Folch J, Lees M, Sloane Stanley GH. A simple method for the isolation and purification of total lipids from animal tissues. J 247. 248. 249. magnetic resonance spectroscopy. J Lipid Res 1996;37:1529-55. Biol Chem 1957;226:497-509. Casu M., Lai A., Pilia A., Casti G., Zedda S., Gibbons W.A. One- and two-dimensional 1H-NMR analysis of lipids extracted from erythrocytes and plasma of humans. Arch Gerontol Geriatr supp 1992;3:111-22. Ruan B, Wilson WK, Pang J, Gerst N, Pinkerton FD, Tsai J, Kelley RI, Whitby FG, Milewicz DM, Garbern J, Schroepfer GJJ. Sterols in blood of normal and Smith-Lemli-Opitz subjects. J Lipid Res 2001;42:799-812. Xiong Q, Ruan B, Whitby FG, Tuohy RP, Belanger TL, Kelley RI, Wilson WK, Schroepfer GJ. A colorimetric assay for 7dehydrocholesterol with potential application to screening for Smith-Lemli-Opitz syndrome. Chem Phys Lipids 2002;115:1-15. 250. 251. 252. 253. Nelson G. Extraction of lipids and purification of extracts. In: Nelson GJ, ed. Blood Lipids and Lipoproteins: Quantitation, Composition and Metabolism. New-York: John Wiley and Sons, 1972:10-24. Kriat M, Vion Dury J, Confort Gouny S, Favre R, Viout P, Sciaky M, Sari H, Cozzone PJ. Analysis of plasma lipids by NMR spectroscopy: application to modifications induced by malignant tumors. J Lipid Res 1993;34:1009-19. Willker W, Leibfritz D. Assignment of mono- and polyunsaturated fatty acids in lipids of tissues and body fluids. Magn Reson Chem 1998;36:S79-S84. Sze DY, Jardetzky O. Characterization of lipid composition in stimulated human lymphocytes by 1H-NMR. Biochim Biophys Acta 1990;1054:198-206. 254. Jira PE, de Jong JG, Janssen Zijlstra FS, Wendel U, Wevers RA. Pitfalls in measuring plasma cholesterol in the Smith-Lemli- 255. Allain CC, Poon LS, Chan CS, Richmond W, Fu PC. Enzymatic determination of total serum cholesterol. Clin Chem Opitz syndrome. Clin Chem 1997;43:129-33. 1974;20:470-5. 207 Reference list 256. 257. Roeschlau P, Bernt E, Gruber W. Enzymatic determination of total cholesterol in serum. Z Klin Chem Klin Biochem 1974;12:226. Passing H, Bablock W. A new biometrical procedure for testing the equality of measurements from two different analytical methods. Application of linear regression procedures for method comparison studies in clinical chemistry, Part I. J Clin Chem Clin Biochem 1983;21:709-20. 258. 259. Tint GS, Irons M, Elias ER, Batta AK, Frieden R, Chen TS, Salen G. Defective cholesterol biosynthesis associated with the Smith-Lemli-Opitz syndrome. N Engl J Med 1994;330:107-13. Björkhem I, Muri Boberg K, Leitersdorf E. Inborn errors in bile acid biosynthesis and storage of sterols other than cholesterol. In: Scriver SC, Beaudet AL, Sly WS, Valle DV, eds. The metabolic & molecular bases of inherited disease. New York: McGraw Hill, 2001:2961-88. 260. Berge KE, Tian H, Graf GA, Yu L, Grishin NV, Schultz J, Kwiterovich P, Shan B, Barnes R, Hobbs HH. Accumulation of dietary cholesterol in sitosterolemia caused by mutations in adjacent ABC transporters. Science 2000;290:1771-5. 261. Salen G, Shefer S, Nguyen L, Ness GC, Tint GS, Shore V. Sitosterolemia. J Lipid Res 1992;33:945-55. 262. Jansen GA, Wanders RJ, De Groot CJ, Moser HW, Mihalik SJ, Watkins PA. Phytanoyl-CoA hydroxylase deficiency: discovery 263. Klenk E, Kahlke W. Über das Vorkommen der 3,7,11,15-Tetramethylhexadecansaure (phytansaure) in den of the long-sought enzyme defect in Refsum disease. N Engl J Med 1996;337:133-4. Cholesterinestern und anderen Lipidfraktionen der Organe bei einem Krankheitsfall unbekanter Genese (Verdacht auf Heredopathia atactica polyneuritiformis, Refsum's syndrome). Hoppe-Seyler Zeits Physiol Chem 1963;333:133-9. 264. Wanders RJ, Jakobs C, Skjeldal OH. Refsum disease. In: Scriver SC, Beaudet AL, Sly WS, Valle DV, eds. The metabolic & 265. Wolthers BG, Walrecht HT, van der Molen JC, Nagel GT, Van Doormaal JJ, Wijnandts PN. Use of determinations of 7- molecular bases of inherited disease. New York: McGraw Hill, 2001:3303-21. lathosterol (5 alpha-cholest-7-en-3 beta-ol) and other cholesterol precursors in serum in the study and treatment of disturbances of sterol metabolism, particularly cerebrotendinous xanthomatosis. J Lipid Res 1991;32:603-12. 266. Styles P, Soffe NF, Scott CA, Cragg DA, Row F, White DJ, White PCJ. A high-resolution NMR probe in which the coil and 267. Keun HC, Beckonert O, Griffin JL, Richter C, Moskau D, Lindon JC, Nicholson JK. Cryogenic probe 13C NMR spectroscopy of preamplifier are cooled with liquid helium. J Magn Reson 1984;60:397-404. urine for metabonomic studies. Anal Chem 2002;74:4588-93. 268. 269. 270. Tanaka K, Armitage IM, Ramsdell HS, Hsia YE, Lipsky SR, Rosenberg LE. [13C]Valine metabolism in methylmalonicacidemia using nuclear magnetic resonance: propinonate as an obligate intermediate. Proc Natl Acad Sci U S A 1975;72:3692-6. Spraul M, Hofmann M, Dvortsak P, Nicholson JK, Wilson ID. Liquid chromatography coupled with high-field proton NMR for profiling human urine for endogenous compounds and drug metabolites. J Pharm Biomed Anal 1992;10:601-5. Pullen FS, Swanson AG, Newman MJ, Richards DS. Online liquid chromatography-nuclear magnetic resonance-mass spectrometry - a powerful spectroscopic tool for the analysis of mixtures pf pharmaceutical interest. Rapid Comm Mass Spectrom 1995;9:1003-6. 271. Lindon JC. HPLC-NMR-MS: past, present and future. Drug Discov Today 2003;8:1021-2. 272. Lindon JC, Nicholson JK, Wilson ID. Direct coupling of chromatographic separations to NMR spectroscopy. Prog Nucl Magn 273. Burton KI, Everett JR, Newman MJ, Pullen FS, Richards DS, Swanson AG. On-line liquid chromatography coupled with high Reson Spec 1996;29:1-49. field NMR and mass spectrometry (LC-NMR-MS): a new technique for drug metabolite structure elucidation. J Pharm Biomed Anal 1997;15:1903-12. 274. Corcoran O, Spraul M. LC-NMR-MS in drug discovery. Drug Discov Today 2003;8:624-31. 275. Spraul M, Freund AS, Nast RE, Withers RS, Maas WE, Corcoran O. Advancing NMR sensitivity for LC-NMR-MS using a 208 cryoflow probe: application to the analysis of acetaminophen metabolites in urine. Anal Chem 2003;75:1536-41. Reference list 276. Peck TL, Magin RL, Lauterbur PC. Design and analysis of microcoils for NMR microscopy. J Magn Reson B 1995;108:114- 277. Olson DL, Peck TL, Webb AG, Magin RL, Sweedler JV. High-resolution microcoil 1H-NMR for mass-limited, nanoliter- 278. Schlotterbeck G, Ross A, Hochstrasser R, Senn H, Kuhn T, Marek D, Schett O. High-resolution capillary tube NMR. A 24. volume samples. Science 1995;270:1967-70. miniaturized 5-microL high-sensitivity TXI probe for mass-limited samples, off-line LC NMR, and HT NMR. Anal Chem 2002;74:4464-71. 279. Griffin JL, Nicholls AW, Keun HC, Mortishire Smith RJ, Nicholson JK, Kuehn T. Metabolic profiling of rodent biological fluids 280. Wolters AM, Jayawickrama DA, Sweedler JV. Microscale NMR. Curr Opin Chem Biol 2002;6:711-6. 281. Liu M, Nicholson JK, Lindon JC. High-resolution diffusion and relaxation edited one- and two-dimensional 1H NMR via 1H NMR spectroscopy using a 1 mm microlitre probe. Analyst 2002;127:582-4. spectroscopy of biological fluids. Anal Chem 1996;68:3370-6. 282. de Graaf RA, Behar KL. Quantitative 1H NMR spectroscopy of blood plasma metabolites. Anal Chem 2003;75:2100-4. 283. Liu M, Tang H, Nicholson JK, Lindon JC. Use of 1H NMR-determined diffusion coefficients to characterize lipoprotein 284. 285. fractions in human blood plasma. Magn Reson Chem 2002;40:S83-S88. Liu M, Nicholson JK, Parkinson JA, Lindon JC. Measurement of biomolecular diffusion coefficients in blood plasma using two-dimensional 1H-1H diffusion-edited total-correlation NMR spectroscopy. Anal Chem 1997;69:1504-9. Beckwith Hall BM, Thompson NA, Nicholson JK, Lindon JC, Holmes E. A metabonomic investigation of hepatotoxicity using diffusion-edited 1H NMR spectroscopy of blood serum. Analyst 2003;128:814-8. 286. 287. Spraul M, Eichhoff U. NMR analysis of mixtures using hyphenation techniques and software. Russian Chem Bull Int Ed 2003;52:2529-38. Wishart DS, Querengesser LMM, Lefebvre BA, Epstein NA, Greiner R, Newton JB. Magnetic Resonance Diagnostics: A new technology for high-throughput clinical diagnostics. Clin Chem 2001;47:1918-21. 288. Dayhoff JE, DeLeo JM. Artificial neural networks: opening the black box. Cancer 2001;91:1615-35. 209 210 Dankwoord Dankwoord Ik wil graag iedereen bedanken die op welke wijze dan ook heeft bijgedragen aan de totstandkoming van dit proefschrift. Mijn promotor, prof. R. Wevers, drijvende kracht achter dit NMR onderzoek. Beste Ron, ik heb veel aan je te danken; jij hebt mij de mogelijkheid gegeven om dit promotieonderzoek uit te voeren. Heel veel dank voor je vasthoudendheid, je vertrouwen en je begeleiding! Ik hoop dat we nog lang zullen samenwerken, met ooit een eigen magneet binnen ons laboratorium. Mijn copromotor, dr. M. Willemsen. Beste Michèl, bedankt voor de inspirerende avonden bij jou thuis waar jij, Ron en ik ideeën en de voortgang van mijn promotie bespraken. Als kinderarts toon je altijd veel belangstelling voor het biochemische onderzoek. Jouw enthousiasme motiveert enorm! En verder een ‘dankjewel’ aan: mijn paranimfen Fokje en Leon, alle coauteurs, vooral Marlies Oostendorp en Marinette van der Graaf, mijn collega’s van het laboratorium Kindergeneeskunde en Neurologie - met een speciale dank aan Sandra Hoenderop -, Jos Joordens, Esther Hermans, Bruker Germany (Manfred Spraul, Hartmut Schäfer and Peter Neidig), Bruker Switzerland (Detlef Moskau and Sandra Loss), alle studenten en arts-assistenten die bij mij stage hebben gelopen. Een bijzonder ‘dankjewel’ aan mijn familie. Mijn schoonouders. Ad en Nelly Hulsken, jullie hebben altijd veel belangstelling voor mijn werkzaamheden in het ziekenhuis. Bedankt voor alles wat jullie voor ons hebben gedaan en doen; als ouders en als opa en oma van Dagmar en Kai. Mijn ouders. Manfred en Marian Engelke. Pa en ma, jullie hebben me altijd de vrijheid gegeven in de keuzes die ik maakte. Bedankt voor dit vertrouwen en voor alles wat jullie voor ons hebben gedaan en doen. Ik kan altijd op jullie rekenen! Robert, bedankt voor de gave omslag van dit boek! En ten slotte dank aan de drie belangrijkste mensen in mijn leven: Astrid, Dagmar en Kai. Lieve Astrid, bedankt voor alles wat was, is en wat nog komen zal … ik hou van je! 211 212 Curriculum Vitae Curriculum Vitae Udo Engelke wordt op 28 april geboren te Kevelaer in Duitsland. Op 7-jarige leeftijd verhuist hij met zijn ouders, broertje Robert en zusje Corina naar Nederland. Hij groeit op in Venray waar hij in 1983 het Mavo-diploma behaalt aan de katholieke Mavo “de Kolk”. Datzelfde jaar begint hij aan de opleiding Middelbaar Laboratoriumonderwijs (MLO) te Venlo die hij in 1987 succesvol afrondt. In 1987 start Udo met de opleiding Hoger Laboratoriumonderwijs (HLO; differentiatie chemie) aan de Hogeschool Eindhoven. Op 23 januari 1992 studeert hij af. Sinds 1 februari 1992 is Udo werkzaam als research-analist op de afdeling laboratorium Kindergeneeskunde & Neurologie van het Universitair Medisch Centrum Nijmegen St Radboud. Het aandachtgebied van Udo is NMRspectroscopie van lichaamsvloeistoffen bij patiënten met erfelijke stofwisselingsziekten. Op 19 december 1997 trouwt Udo met Astrid Hulsken. Twee jaar later verhuizen ze vanuit Nijmegen naar Molenhoek, waar ze samen met hun kinderen, Dagmar (1999) en Kai (2003) nog steeds wonen. 213 214 List of publications List of publications • Engelke UF, Sass JO, van Coster RN, Gerlo E, Olbrich H, Krywawych S, Calvin J, Hart C, Omran H and Wevers RA NMR Spectroscopy of aminoacylase 1 deficiency, a novel inborn error of metabolism NMR Biomed 2007: in press • Willemsen MA, Engelke UF, van der Graaf M and Wevers RA Methylsulfonylmethane (MSM) ingestion causes a significant resonance in proton magnetic resonance spectra of brain and cerebrospinal fluid Neuropediatrics 2006; 37: 312-314 • Engelke UF, Moolenaar SH, Hoenderop SM, Morava E, van der Graaf M, Heerschap A and Wevers RA Handbook of 1H-NMR spectroscopy in inborn errors of metabolism: body fluid NMR spectroscopy and in vivo MR spectroscopy 2nd edition; 2006 SPS Verlagsgesellschaft mbH, Heilbronn, In press • Engelke UF, Oostendorp M, Wevers RA NMR spectroscopy of body fluids as a metabolomics approach to inborn errors of metabolism IN: The Handbook of Metabonomics and Metabolomics 2007: (14) 373 – 410 Editors: John C. Lindon, Jeremy K. Nicholson and Elaine Holmes Published by Elsevier B.V. • Burlina AP, Schmitt B, Engelke UF, Wevers RA, Burlina AB, Boltshauser E. Hypoacetylaspartia: clinical and biochemical follow-up of a patient. Adv Exp Med Biol. 2006; 576:283-7; discussion 361-3. • Engelke UF, Oostendorp M , Willemsen MA, Wevers RA. Diagnosing inborn errors of lipid metabolism with proton nuclear magnetic resonance spectroscopy. Clin Chem. 2006;52(7):1395-405. 215 List of publications • Engelke UF, Kremer B, Kluijtmans LA, van der Graaf M, Morava E, Loupatty FJ, Wanders RJ, Moskau D, Loss S, van den Bergh E, Wevers RA. NMR spectroscopic studies on the late onset form of 3-methylglutaconic aciduria type I and other defects in leucine metabolism. NMR Biomed. 2006; 19(2): 271-8. • Wortmann S, Rodenburg RJ, Huizing M, Loupatty FJ, de Koning T, Kluijtmans LA, Engelke UF, Wevers R, Smeitink JA, Morava E. Association of 3-methylglutaconic aciduria with sensori-neural deafness, encephalopathy, and Leigh-like syndrome (MEGDEL association) in four patients with a disorder of the oxidative phosphorylation. Mol Genet Metab. 2006; 88(1):47-52. • Sass JO, Mohr V, Olbrich H, Engelke UF, Horvath J, Fliegauf M, Loges NT, SchweitzerKrantz S, Moebus R, Weiler P, Kispert A, Superti-Furga A, Wevers RA, Omran H. Mutations in ACY1, the gene encoding aminoacylase 1, cause a novel inborn error of metabolism. Am J Hum Genet. 2006;78(3):401-9. • Van Coster RN, Gerlo EA, Giardina TG, Engelke UF, Smet JE, De Praeter CM, Meersschaut VA, De Meirleir LJ, Seneca SH, Devreese B, Leroy JG, Herga S, Perrier JP, Wevers RA, Lissens W. Aminoacylase I deficiency: a novel inborn error of metabolism. Biochem Biophys Res Commun. 2005;338(3):1322-6. • Engelke UF, Tangerman A, Willemsen MA, Moskau D, Loss S, Mudd SH, Wevers RA. Dimethyl sulfone in human cerebrospinal fluid and blood plasma confirmed by one- dimensional (1)H and two-dimensional (1)H-(13)C NMR. NMR Biomed. 2005;18(5):331-6. • Boltshauser E, Schmitt B, Wevers RA, Engelke UF, Burlina AB, Burlina AP. Follow-up of a child with hypoacetylaspartia. Neuropediatrics. 2004;35(4):255-8. • Wolf NI, Haas D, Hoffmann GF, Jakobs C, Salomons GS, Wevers RA, Engelke UF, Rating D. Sedation with 4-hydroxybutyric acid: a potential pitfall in the diagnosis of SSADH deficiency. J Inherit Metab Dis. 2004;27(2):291-3. 216 List of publications • Wolf NI, Willemsen MA, Engelke UF, van der Knaap MS, Pouwels PJ, Harting I, Zschocke J, Sistermans EA, Rating D, Wevers RA. Severe hypomyelination associated with increased levels of N-acetylaspartylglutamate in CSF. Neurology. 2004;62(9):1503-8. • Groenen PM, Engelke UF, Wevers RA, Hendriks JC, Eskes TK, Merkus HM, SteegersTheunissen RP. High-resolution 1H NMR spectroscopy of amniotic fluids from spina bifida fetuses and controls. Eur J Obstet Gynecol Reprod Biol. 2004;112(1):16-23. • Engelke UF, Liebrand-van Sambeek ML, de Jong JG, Leroy JG, Morava E, Smeitink JA, Wevers RA. N-acetylated metabolites in urine: proton nuclear magnetic resonance spectroscopic study on patients with inborn errors of metabolism. Clin Chem. 2004;50(1):58-66. • Mazon Ramos A, Gil-Setas A, Berrade Zubiri S, Bandres Echeverri T, Wevers R, Engelke UF, Zschocke J. Primary trimethylaminuria or fish odor syndrome. A novel mutation in the first documented case in Spain Med Clin (Barc). 2003;120(6):219-21. Spanish. • Wevers R.A., Engelke U.F.H., Moolenaar, S.H. Proton NMR spectroscopy on body fluids In: Physician’s Guide to the Laboratory Diagnosis of Metabolic Diseases Editors: Blau N, Duran M, Blaskovics ME, Gibson KM. Springer-Verlag Berlin Heidelberg 2003 • Morava E, Kosztolanyi G, Engelke UF, Wevers RA. Reversal of clinical symptoms and radiographic abnormalities with protein restriction and ascorbic acid in alkaptonuria. Ann Clin Biochem. 2003;40(Pt 1):108-11. • Ruitenbeek W, Kobayashi K, Iijima M, Smeitink JA, Engelke UF, De Abreu RA, Kwast HT, Saheki T, Boelen CA, De Jong JG, Wevers RA. Moderate citrullinaemia without hyperammonaemia in a child with mutated and deficient argininosuccinate synthetase. Ann Clin Biochem. 2003;40(Pt 1):102-7. 217 List of publications • Moolenaar SH, Engelke UF, Wevers RA. Proton nuclear magnetic resonance spectroscopy of body fluids in the field of inborn errors of metabolism. Ann Clin Biochem. 2003;40(Pt 1):16-24. Review. • Van Cappellen Van Walsum AM, Jongsma HW, Wevers RA, Nijhuis JG, Crevels J, Engelke UF, De Abreu RA, Moolenaar SH, Oeseburg B, Nijland R. 1H-NMR spectroscopy of cerebrospinal fluid of fetal sheep during hypoxia-induced acidemia and recovery. Pediatr Res. 2002;52(1):56-63. • Moolenaar SH, Engelke UF, Abeling NG, Mandel H, Duran M, Wevers RA. Prolidase deficiency diagnosed by 1H NMR spectroscopy of urine. J Inherit Metab Dis. 2001;24(8):843-50. • Moolenaar SH, Göhlich-Ratmann G, Engelke UF, Spraul M, Humpfer E, Dvortsak P, Voit T, Hoffmann GF, Bräutigam C, van Kuilenburg AB, van Gennip A, Vreken P, Wevers RA. beta-Ureidopropionase deficiency: a novel inborn error of metabolism discovered using NMR spectroscopy on urine. Magn Reson Med. 2001;46(5):1014-7. • Moolenaar SH, van der Knaap MS, Engelke UF, Pouwels PJ, Janssen-Zijlstra FS, Verhoeven NM, Jakobs C, Wevers RA. In vivo and in vitro NMR spectroscopy reveal a putative novel inborn error involving polyol metabolism. NMR Biomed. 2001;14(3):167-76. • van Cappellen van Walsum AM, Jongsma HW, Wevers RA, Nijhuis JG, Crevels J, Engelke UF, Moolenaar SH, Oeseburg B, Nijland R. Hypoxia in fetal lambs: a study with (1)H-MNR spectroscopy of cerebrospinal fluid. Pediatr Res. 2001;49(5):698-704. • Binzak BA, Wevers RA, Moolenaar SH, Lee YM, Hwu WL, Poggi-Bach J, Engelke UF, Hoard HM, Vockley JG, Vockley J. Cloning of dimethylglycine dehydrogenase and a new human inborn error of metabolism, dimethylglycine dehydrogenase deficiency. Am J Hum Genet. 2001;68(4):839-47. 218 List of publications • Boss EA, Moolenaar SH, Massuger LF, Boonstra H, Engelke UF, de Jong JG, Wevers RA. High-resolution proton nuclear magnetic resonance spectroscopy of ovarian cyst fluid. NMR Biomed. 2000;13(5):297-305. • van der Knaap MS, Wevers RA, Struys EA, Verhoeven NM, Pouwels PJ, Engelke UF, Feikema W, Valk J, Jakobs C. Leukoencephalopathy associated with a disturbance in the metabolism of polyols. Ann Neurol. 1999;46(6):925-8. • Lehnert W, Stogmann W, Engelke UF, Wevers RA, van den Berg GB. Long-term follow up of a new case of hawkinsinuria. Eur J Pediatr. 1999;158(7):578-82. • Wevers RA, Engelke UF, Moolenaar SH, Bräutigam C, de Jong JG, Duran R, de Abreu RA, van Gennip AH. 1H-NMR spectroscopy of body fluids: inborn errors of purine and pyrimidine metabolism. Clin Chem. 1999;45(4):539-48. • Moolenaar SH, Poggi-Bach J, Engelke UF, Corstiaensen JM, Heerschap A, de Jong JG, Binzak BA, Vockley J, Wevers RA. Defect in dimethylglycine dehydrogenase, a new inborn error of metabolism: NMR spectroscopy study. Clin Chem. 1999;45(4):459-64. • Massuger LF, van Vierzen PB, Engelke UF, Heerschap A, Wevers R. 1H-magnetic resonance spectroscopy: a new technique to discriminate benign from malignant ovarian tumors. Cancer. 1998;82(9):1726-30. • Wevers RA, Engelke UF, Rotteveel JJ, Heerschap A, De Jong JG, Abeling NG, van Gennip AH, de Abreu RA. 1H NMR spectroscopy of body fluids in patients with inborn errors of purine and pyrimidine metabolism. J Inherit Metab Dis. 1997;20(3):345-50. • Wevers RA, Engelke UF, Wendel U, de Jong JG, Gabreëls FJ, Heerschap A. Standardized method for high-resolution 1H-NMR of cerebrospinal fluid. Clin Chem. 1995;41(5):744-51. 219 List of publications • Abeling NG, van Gennip AH, Bakker HD, Heerschap A, Engelke UF, Wevers RA. Diagnosis of a new case of trimethylaminuria using direct proton NMR spectroscopy of urine. J Inherit Metab Dis. 1995;18(2):182-4. • Wevers RA, Engelke UF, Heerschap A. High-resolution 1H-NMR spectroscopy of blood plasma for metabolic studies. Clin Chem. 1994;40(7 Pt 1):1245-50. 220