Metabonomic and Metallomic Profiling in the Amniotic Fluid of

Mar 19, 2008 - Received November 21, 2007. Abstract: Epidemiology and studies in animal models have revealed that prenatal malnutrition is highly cor-...
1 downloads 0 Views 2MB Size
Download PDF
Metabonomic and Metallomic Profiling in the Amniotic Fluid of Malnourished Pregnant Rats Qing Shen,‡,# Xin Li,†,# Yunping Qiu,† Mingming Su,† Yumin Liu,† Houkai Li,† Xiaoyan Wang,† Xiangyu Zou,§ Chonghuai Yan,§ Lan Yu,‡ Sheng Li,‡ Chunling Wan,| Lin He,*,‡ and Wei Jia*,† School of Pharmacy, Shanghai Jiao Tong University, Shanghai, 200240, People’s Republic of China, Institute for Nutritional Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, 200031, People’s Republic of China, Shanghai Xinhua Hospital affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai, 200092, People’s Republic of China, and Bio-X Center, Shanghai Jiao Tong University, Shanghai, 200030, People’s Republic of China Received November 21, 2007

Abstract: Epidemiology and studies in animal models have revealed that prenatal malnutrition is highly correlated with abnormal fetal neurodevelopment. We present here a combined metabonomic and metallomic profiling technique to associate the metabolic and trace-elemental composition variations of rat amniotic fluid (AF) in malnourished pregnant rats with the retardation of fetal rat neurodevelopment. The AF samples from three groups of pregnant Sprague–Dawley rats, which were fed either a normal diet, a low-protein diet, or “a famine diet”, were subjected to GC/MS and ICP/MS combined with multivariate data analysis (MVDA). PCA scores plot of both GC/ MS and ICP/MS data showed similar and unique metabolic signatures of AF in response to the different diets. Rats in the famine group released increased amounts of glycine, inositol, putrescine, and rubidium and decreased amounts of methionine, dopa, tryptophan, glutamine, zinc, cobalt, and selenium in the AF. These discriminable variations in the AF may indicate the abnormality of a number of metabolic pathways in fetal rats including the folate cycle and methionine pathway, the monoamine pathway, and tri-iodothyronine (T3) metabolism. The abnormalities may be the result of metabolites or elemental differences or a combination of both. This study demonstrates the potential of combining profiling of small-molecule metabolites and trace elements to broaden the understanding of biological variations associated with fetal neurodevelopment induced by environmental perturbation. * Corresponding authors. Prof. Wei Jia, School of Pharmacy, Shanghai Jiao Tong University, 800 Dongchuan Road, Shanghai, 200240, China. Tel.: +86 21 62932292. Fax: +86 21 62932292. E-mail: [email protected]. Prof. Lin He, Institute for Nutritional Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, 294 Taiyuan Road, Shanghai, 200031, China. Tel./fax: +86 21 62822491. E-mail: [email protected]. † School of Pharmacy, Shanghai Jiao Tong University. ‡ Institute for Nutritional Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences. § Shanghai Xinhua Hospital affiliated to Shanghai Jiao Tong University School of Medicine. | Bio-X Center, Shanghai Jiao Tong University. # These authors contributed equally to this work. 10.1021/pr700776c CCC: $40.75

 2008 American Chemical Society

Keywords: metabonomics • metallomics • neurodevelopment • metabolic profiling • trace elements • prenatal malnutrition • amniotic fluid

Introduction Many complex interactive factors such as genes, environment, lifestyle, and diet (nutrition) determine fetal development in the neurons, nerves, synapses, and brain.1 Nutrition is probably the single greatest environmental influence on the fetus, and it plays a necessary role in the maturation and functional development of the fetal central nervous system.2 Conventional studies focus mainly on abnormal genetic and epidemiological outcomes of fetal neurodevelopment caused by maternal malnutrition.3 Understanding the pathological effects of prenatal malnutrition will undoubtedly enrich current knowledge of fetal development and provide insight into the neurodevelopmental process. However, the in-depth study of impairment of the fetal nervous system during pregnancy is a tough task primarily due to limitations in analytical methods. Recent advances4 in the incorporation of genomic, proteomic, metabonomic, and metallomic bioinformation into a holistic biological network may provide comprehensive insight to understanding the mechanism of abnormal fetal neurodevelopment due to prenatal malnutrition. Metabonomics is a new technology that applies advanced separation and detection methods to capture the metabolome, the collection of small molecules that characterizes metabolic pathways. Metabonomics is defined as “the quantitative measurement of the dynamic multiparametric metabolic response of living systems to pathophysiological stimuli or genetic modification”,5 and this approach attempts to capture global change and overall physiological status in biochemical networks and pathways to elucidate specific sites of perturbations.6 The development of mass spectrometry (MS)-based metabonomic technology has increased the number of identified metabolites, which offers more opportunities to uncover the mechanisms of diverse diseases.7,8 Various elements have been shown to play critical roles in biological systems.9 Most of these trace elements, especially trace metals, in biological fluids and organs are binding with various proteins, forming “metalloproteins” or “metalloenzymes” that are essential in regulating biological reactions and physiological functions in cells and organs. In the area of Journal of Proteome Research 2008, 7, 2151–2157 2151 Published on Web 03/19/2008

technical notes

Shen et al. compare global metabolic and elemental changes in AF samples from rat models fed either a low-protein diet or a normal diet. The purpose of this study was to characterize the metabolic and elemental signature variations caused by dietary alteration and to determine the association between prenatal malnutrition and potential hazards to fetal neurodevelopment.

Experimental Methods

Figure 1. Typical GC/MS total ion current (TIC) chromatograms of amniotic fluid of pregnant rats: (A) the control group, (B) the low-protein group, or (C) the famine group. The differentially expressed metabolites, each labeled with a number, are provided in Table 2.

metallomics, the most important research target is elucidation of the physiological or pathological roles and functions of biomolecules binding with metallic ions in the biological systems.10 The emergence of element-detection techniques, such as inductively coupled plasma atomic emission spectrometry (ICP/AES) and inductively coupled plasma mass spectrometry (ICP/MS)-based metallomics, enables us to capture metabolic variation due to trace element perturbation in an organism.11 Amniotic fluid (AF) is the unique biofluid of mammals during pregnancy. It is generally believed that the fetus exchanges nutrients and metabolites with the maternal environment through amniotic fluid as well as blood in a dynamic process.12 The endogenous metabolites present in both biofluids not only reflect parental nutritional status but also exert a significant influence on fetal development and growth as key components of the maternal environment. In this study, we used modern analytical instruments, including GC/MS and ICP/MS, combined with multivariate data analysis (MVDA) to map and 2152

Journal of Proteome Research • Vol. 7, No. 5, 2008

Animal Handling Procedure and Sampling. A total of 18 female Sprague–Dawley rats (weighing 200 ( 10 g) were purchased from Shanghai Laboratory Animal Co. Ltd. (SLAC, Shanghai, China) and housed under a controlled condition of 12 h light/12 h dark cycle at 23 ( 2 °C and 35 ( 5% humidity. Following a two-week accommodation period, each rat was mated and day 0 (the day of pregnancy) was defined by a sperm-positive vaginal smear or vaginal plug.13 The pregnant animals were divided randomly into three groups as follows (n ) 6 for each group): the control group, given a standard diet (D12450B); the low-protein group, dosed with the lowprotein diet (casein) (D06022301); and the famine group, fed with half of the diet supplied for the low-protein group (Table S1). All the rats had access to water ad libitum. Animals were sacrificed on the 19th day of pregnancy (weight of both maternal and fetal rats was measured after sacrifice), and AF samples were drawn from each uterus and immediately stored at -80 °C. Experimental procedures and protocols were approved by the Institutional Animal Care and Use Committee at the Institute for Nutritional Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences. AF Sample Preparation and Derivatization. AF samples from the three groups were randomly selected for GC/MS sample preparation and data acquisition to eliminate systematic bias. Each 200 µL aliquot of AF was added into a 1.5 mL of tube following the addition of 400 µL of acetone for protein precipitation. The mixture was vortexed for 30 s and centrifuged at 10 000 rpm for 10 min. A total of 400 µL of aqueous supernatant was transferred to a 500 µL glass tube and dried under vacuum. The dried analytes were derivatized with 80 µL of methylhydroxylamine hydrochloride (15 mg/mL, dissolved in pyridine) for 90 min at 37 °C and silylated for 2 h at 70 °C with 80 µL of bis(trimethylsilyl)-trifluoroacetamide (BSTFA) (containing 1% trimethylchlorosilane). Each 70 µL aliquot of hexane was added to the derivatization bottles. After being vortexed for 1 min and incubated at room temperature for an hour, the derivatized mixtures were subjected to GC/MS analysis.14 GC/MS Spectral Acquisition. The derivatized AF samples were analyzed using a PerkinElmer gas chromatograph coupled with a TurboMass-Autosystem XL mass spectrometer (PerkinElmer Inc., Waltham, MA, USA). Each 1 µL aliquot of the extracts was injected into a DB-5MS capillary column coated with 5% Diphenyl cross-linked 95% dimethylpolysiloxane (30 m × 250 µm i.d., 0.25 µm film thickness; Agilent J&W Scientific, Folsom, CA, USA) in the splitless mode. Both the injection temperature and the interface temperature were set to 260 °C, and the ion source temperature was adjusted to 200 °C. Initial GC oven temperature was set at 80 °C, and 2 min after injection, the GC oven temperature was increased to 285 °C at a rate of 5 °C/min and held at 285 °C for 7 min. Helium was used as the carrier gas with a flow rate of 1 mL/min. Measurements were made with electron impact ionization (70 eV) in the full scan mode (m/z 30–550).15 All the experiments were conducted

Metabonomic and Metallomic Profiling

technical notes

Figure 2. PCA scores plot (PC1/PC2) of (A) GC/MS data and (B) ICP/MS data derived from the amniotic fluid of rats from the three different diet groups (blue, the control group; black, the low-protein group; red, the famine group). Each dot denotes one rat.

Figure 3. Heat map showing differences in individual metabolites and trace elements detected from the AF in the diet-deficiency group rats as compared to controls. LP/C: low-protein group versus control group; F/C: famine group versus control group. The different metabolites/elements in each cell whose fold-change level was higher in nutritional deficiency groups versus the control group are shown in red, while those with decreased levels are shown in green. The critical p value of the M-W test was 0.05.

according a to fully randomized run in order to avoid any artificial variance. AF Sample Preparation for ICP/MS Analysis. Each 0.2 g of the AF sample was added dropwise to the Teflon vessels and digested with 400 µL of concentrated nitric acid in a microwave digestion system, as detailed in Table S2. After digestion, nitric acid (0.1%, Fisher TraceMetal grade) was used to dilute each sample to about 5 mL. The diluted samples were stored at 4 °C pending ICP/MS analysis.16 ICP/MS Analytical Instruments and Conditions. An Agilent 7500a ICP/MS system (Agilent, Tokyo, Japan) was used with the following parameters: detection range was 2–260 amu; rf forward power was 1200 W; sample depth was 6.9 mm; carrier gas (Ar, argon) flow was 1.12 L/min; extract 1 and extract 2 were -111 and -25 V. All parameters were optimized to the highest ratio of signal-to-noise using the following internal standards: 6Li, 45Sc, 72Ge, 89Y, 115In, 159Tb, and 209Bi. Data Process. All GC/MS raw data files were converted to CDF format via Databridge (Perkin-Elmer Inc., Waltham, MA,USA) and directly processed by our custom scripts in MATLAB 7.0 (The MathWorks, Inc. Natick, MA, USA), where baseline correction, peak deconvolution and alignment, internal standard exclusion, and normalization to the total sum of the chromatogram were carried out. The resulting threedimensional matrix consisting of an arbitrary compound index (paired retention time-m/z), sample names (observations), and normalized peak areas (variables) was imported into the SIMCA-P 11.0 Software package (Umetrics, Umeå, Sweden) for MVDA. Similarly, each ICP/MS raw data file was exported through the Agilent 7500a operation system and organized in a spreadsheet with the form of element index, sample names (observations), and signal response (variables). Both GC/MS

and ICP/MS data were autoscaled (mean-centered and scaled to unit variance) using principal component analysis (PCA) to visualize general clustering, trends, or outliers among the observations. To obtain the differentially expressed metabolites responsible for the separation between groups, a more sophisticated partial least-squares-discriminant analysis (PLS-DA) model was utilized on the GC/MS and ICP/MS data sets. To validate the model and avoid overfitting, a six-round crossvalidation was carried out with 1/6th of the samples being excluded from the model in each round. This procedure was repeated in an iterative manner until each sample had been excluded once. The differential expression metabolites were attained using a cutoff threshold of correlation coefficients (r > 0.5 was considered as having good discriminant ability). Additionally, heat maps (Figure 3) were created to visualize the metabolites and elemental concentration variation induced by the different diets. Each cell in the heat map represents a fold change between two groups (e.g., the famine group or the lowprotein group versus the control group) for a particular metabolite. Based on thresholds of the p values of the nonparametric Mann–Whitney (M-W) test implemented in the Matlab statistical toolbox (the critical p value was set to 0.05), metabolites significantly increased due to diet were coded with red, while metabolites significantly decreased due to diet were displayed in green. Each color corresponded to the magnitude of the difference. Ultimately, the differentially expressed metabolites were obtained by both the cutoff threshold of correlation coefficients from a cross-validated PLS-DA model and the critical p values from the M-W test to avoid potentially spurious associations in terms of the borderline number of samples per group. Journal of Proteome Research • Vol. 7, No. 5, 2008 2153

technical notes

Shen et al.

Table 1. Maternal Weight Percentage Gain (WPG) and the Fetal Weight in Different Diet Intervention Groups group

maternal WPG (%)

fetal weight (g)

control low protein famine

51.00 ( 8.84 50.07 ( 5.46 13.65 ( 7.21b

2.97 ( 0.61 2.66 ( 0.16 2.25 ( 0.36a

a P < 0.05, compared with the control group. with the control group.

b

P < 0.01, compared

Results Maternal and Fetal Weight. Maternal and fetal weights are shown in Table 1. Both maternal weight gain and fetal weight were lower in the famine group compared to the control group, while the weights in the low-protein diet group were not statistically different from the control group. Metabolic and Trace Elemental Profiling Analysis of AF Samples. The typical GC/MS total ion current (TIC) chromatograms obtained from AF samples are illustrated in Figure 1. Based on available libraries, including NIST, Wiley, and NBS, 98 metabolites (approximately 65% of all detected peaks) were identified including amino acids, fatty acids, carbohydrates, and polyamines, etc. (Table S3). As shown in Figure 1, differences in the metabolite spectra were readily observed between the three groups. For instance, an increase of inositol and a decrease of methionine were observed in the famine group when compared with the control group. However, it is not feasible to visually elucidate the global biochemical variation of the metabolome from a GC/MS TIC chromatogram. Therefore, PCA was used to reduce the megavariate data into the low-dimensional plane so as to visualize variations of metabolic expression patterns. The PCA scores plot (PC1 versus PC2) derived from GC/MS data (Figure 2A) identified three distinct clusters representing the control group, the low-protein group, and the famine group. The clear separation of the control group from the two malnutrition groups occurred in PC1, whereas separation of the famine group from the low-protein group occurred in PC2. Hence, the PCA scores plot of GC/MS data identified unique metabolic signatures of AF under conditions of prenatal malnutrition. In parallel, profiling of trace elements present in the AF was carried out using an analytical procedure with ICP/MS, in which 65 elements were consistently detected across all the samples. The use of internal standards (6Li, 45Sc, 72Ge, 89Y, 115In, 159 Tb, and 209Bi) and dilution of the AF were necessary for minimizing signal drift and matrix effects. Multivariate and univariate statistical analyses were performed, and three distinct clusters corresponding to the three different diets were visualized on the PCA scores plot (Figure 2B). PCA of trace elements in AF samples showed a separation tendency similar to that of the small-molecule metabolites. Metabolites and Trace Elements Involved in Neurodevelopment. On the basis of the correlation coefficients17 of the PLS-DA model and p values of the M-W test utilized in the study, a subset of differentially expressed metabolites in AF corresponding to the three different diets were identified by comparing the major ion fragments of the compound detected with those from commercially available and in-house libraries (Table 2). For example, the levels of putrescine, cadaverine, and glycine were higher, and the levels of methionine, dopa, and gluconate were lower in the AF of the famine group as compared to the control group. Similarly, a list of differentially expressed trace elements in the AF, such 2154

Journal of Proteome Research • Vol. 7, No. 5, 2008

as a higher level of rubidium and a lower level of zinc in the famine group (Table 3), were identified. A heat map was prepared (Figure 3) to illustrate these differentially expressed small-molecule metabolites and trace elements in the AF of rats in the famine and low-protein groups and to provide an overview of metabolic alterations in the two groups.

Discussion In this study, AF samples of pregnant Sprague–Dawley rats were analyzed by GC/MS and ICP/MS to compare the global metabolite and trace element levels resulting from different diets. Distinctly different patterns of metabolites and trace elements were observed in malnourished rats. The differentially expressed small-molecule metabolites and trace elements, representing important metabolic regulatory pathways, provide further understanding of prenatal nutritional status and global biochemical perturbation leading to impairment of fetal neurodevelopment. The methionine cycle, which provides one carbon group in many important metabolic syntheses, plays a vital role in prenatal fetal development and growth.18 Our observations from GC/MS-based metabolic analysis of the downregulation of methionine and cystine in the AF samples from the famine group in comparison with that of the control group suggested an altered methionine metabolic pathway in the fetuses induced by the low-casein maternal diet during the pregnancy. Moreover, as the active site in the corrin core of VitB12, which is a cofactor in regulation of methionine syntheses, cobalt (Co) was detected at a lower concentration in the famine group by ICP/MS analysis. In addition, the decreased zinc level in the AF of the famine group may induce a decrease in the absorption of dietary folate.19 When the results of the metabonomic and metallomic analytical methods were combined, a disturbed methionine and folate metabolic pathway became evident (Figure 4A). Dopaminergic neurons are known to migrate and differentiate during the first trimester of gestation,20 establishing connections throughout the cortex by midgestation.21 Prenatal malnutrition results in the enhanced release and turnover of dopamine in the developing fetal nervous system and a decrease in the number of dopamine binding sites in the striatum.22 Since dopamine in the brain is mainly formed from dopa,23 the observed lower concentration of dopa in the AF of the famine group may be indicative of the impaired developmental process of dopaminergic neurons as the result of prenatal malnutrition. Meanwhile, a higher level of rubidium (Rb) was observed in the AF samples of the famine group. This trace element is believed to mimic the role of potassium in adjusting the Na+/K+-stimulated ATPase involved in nerve regulation. A recent study reported a decreased extraneuronal dopamine concentration in the nucleus accumbens of RbCltreated rats, which indicates a condition of reduced dopaminergic neuronal activity in the brain.24 The abnormally higher concentration of Rb, coupled with a lower dopa concentration in the AF of the famine group, may be an indication of the impairment of fetal dopaminergic neuron development under parental malnutrition. Selenium (Se) was found at a significantly lower concentration in the AF samples of the famine group. The importance of selenium in fetal neurodevelopment lies in its participation in metalloenzymes converting the prohormone tetraiodothyronine (T4) to the metabolically active hormone triiodothyronine (T3) (deiodinases I, II, and III).25 Present findings regarding regulatory mechanisms involved in the bioavailability of T3 in

technical notes

Metabonomic and Metallomic Profiling

Table 2. Statistical Analysis of Metabolites Identified by the PCA Model to be Accounted for in the Separation between Famine and Control Groups and the Separation between Low-Protein and Control Groups famine vs control no.

metabolitesa

corr. coeffs.b

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

2-oxo butanoate Hydroxyacetate Sarcosine Glycine Butenedioate Leucine Proline 2,3-dihydroxybutanoate Ethylsulfonate Dopa Methionine Glucitol 2-O-methyl-L-ascorbirate Ribitol 2-keto-D-gluconate Gluconate Glutamine Inositol Uric acid Putrescine Cadaverine Cystine Tryptophane Maltose

0.81 (v) 0.59 (v) 0.65 (v) 0.65 (v) -0.85 (V) -0.68 (V) -0.72 (V) -0.89 (V) -0.72 (V) -0.70 (V) -0.61 (V) 0.69 (v) -0.86 (V) 0.59 (v) 0.58 (v) -0.58 (V) -0.67 (V) 0.88 (v) 0.87 (v) 0.72 (v) 0.74 (v) -0.66 (V) -0.70 (V) -0.58 (V)

low-protein vs control

P P (M-W test) fold changec corr. coeffs. (M-W test) fold change

0.0013 0.0417 0.0215 0.0216 0.0005 0.0149 0.0080 0.0001 0.0085 0.0118 0.0347 0.0123 0.0004 0.0456 0.0500 0.0497 0.0160 0.0001 0.0002 0.0089 0.0058 0.0184 0.0113 0.0465

2.7 1.8 2.3 2.0 -2.7 -2.3 -2.4 -2.7 -2.1 -2.1 -2.0 2.3 -2.7 2.7 1.8 -1.8 -2 2.7 2.5 2.3 2.1 -2.1 -2.7 -1.8

0.67 (v) 0.78 (v) 0.33 (v) 0.70 (v) -0.83 (V) -0.14 (V) -0.82 (V) -0.87 (V) -0.15 (V) -0.75 (V) -0.60 (V) 0.8 (v) 0.09 (V) 0.63 (v) 0.74 (v) 0.07 (V) -0.26 (V) 0.49 (v) 0.82 (v) 0.18 (v) 0.81 (v) -0.30 (V) -0.83 (V) 0.05 (V)

0.0162 0.0031 N.S.d 0.0116 N.S. N.S. 0.0012 0.0003 N.S. 0.0054 0.0381 0.0016 N.S. 0.0297 0.0060 N.S. N.S. N.S. 0.0012 N.S. 0.0014 N.S. 0.0009 N.S.

2.5 2.5 1.4 2.7 -2.4 1.9 -2.5 -2.7 -1.1 -2.4 -1.9 2.5 -1.1 2.7 2.1 -1.2 -1.4 1.9 2.5 1.1 2.7 -1.3 -2.7 -1.1

pathway (KEGG)e

Glycine metabolism Glycine metabolism TCA Amino acid metabolism Amino acid metabolism Methiomine metabolism Tyrosine metabolism Methiomine metabolism Carbohydrate metabolism Vitamin & cofactor metabolism Nucleic acid metabolism Carbohydrate metabolism Carbohydrate metabolism Glutamate metabolism Inositol metabolism Purine metabolism Polyamine metabolism Polyamine metabolism Methiomine metabolism Tryptophane metabolism Carbohydrate metabolism

a Metabolites: the ultimate different expression metabolites were obtained from the overlapped results from a specified correlation coefficient of the PLSDA model and critical P value of the Mann–Whitney (M-W) test. b Corr. coeffs.: Correlation coefficients obtained from a cross-validated PLS-DA model with R2(X) ) 0.87 and Q2(X) ) 0.79, both of which indicate a reliable model. The positive value of correlation coefficients (also depicted with an up arrow) means a higher level of metabolites in the famine group or low-protein group as compared to the control, whereas the negative value of correlation coefficients (also depicted with a down arrow) represents a lower level of metabolites. c Fold change: fold change value for a specific metabolite was calculated by a nonparametric M-W test. d N.S. ) nonsignificant. e (KEGG) ) Kyoto Encyclopedia of Genes and Genomes (http://www.genome.jp/kegg/).

Table 3. Statistical Analysis of Elements Identified by the PCA Model to be Accounted for in the Separation between Famine and Control Groups and the Separation between Low-Protein and Control Groups famine vs control

low-protein vs control

elements

corr. coeffs.

P (M-W test)

fold change

corr. coeffs.

P (M-W test)

fold change

Iron (Fe) Zirconium (Zr) Manganese (Mn) Rubidium (Rb) Copper (Cu) Zinc (Zn) Selenium (Se) Cobalt (Co) Magnesium (Mg) Calcium (Ca) Aluminum (Al)

-0.91 (V) -0.80 (V) -0.90 (V) 0.78 (v) 0.58 (v) -0.58 (V) -0.55 (V) -0.60 (V) -0.62 (V) -0.52 (V) -0.53 (V)

0.0037 0.0037 0.0039 0.0050 0.0247 0.0250 0.0483 0.0480 0.0247 0.0234 0.0483

-2.7 -2.7 -2.7 2.6 2.1 -2.1 -1.9 -1.9 -2.1 -2.1 -1.9

-0.49 (V) -0.82 (V) -0.83 (V) 0.7 (v) 0.56 (v) -0.58 (V) -0.54 (V) -0.45 (V) -0.08 (V) -0.23 (V) -0.55 (V)

N.S. 0.0037 0.0064 0.0129 0.0161 0.0447 0.0447 N.S. N.S. N.S. N.S.

-1.5 -2.7 -2.5 2.3 2.3 -1.9 -1.9 -1.6 1.1 1.3 -1.7

the fetal cortex during early development, as well as the early expression of nuclear thyroid hormone receptors already occupied by T3, strongly support the hypothesis that an adequate supply of maternal T4 is needed for the development of the cerebral cortex during pregnancy.26 The lower expression level of Se in the AF of the famine group may be an indication of an abnormality resulting in T3-mediated interference of fetal brain development. In addition to its influence on folate absorption, zinc (Zn) plays an important role in maintaining the normal enzyme conformation of superoxide dismutase (SOD) and functions of the CNS. Zinc deficiency could result in decreased microtubule polymerization, leading to impairment of brain development and function.27

Using the KEGG database and literature mining, we have also linked fetal neural and brain development changes to some putative important small-molecule metabolites that play a vital role in the corresponding pathways. Some of these metabolites (e.g., putrescine and inositol) that are involved in cell apoptosis28 and the transmembrane signal transduction system29 have already been shown to play an important role in neuronal, synaptic, and brain development (Figure 4C and D).30 Other metabolites (e.g., tryptophan,31,32 glutamine33) that are related to neurotransmitters are known to be involved in many neuron, nerve, and brain developmental events, including neural induction, migration, differentiation, and synaptogenesis (Figure 4B).34 Journal of Proteome Research • Vol. 7, No. 5, 2008 2155

technical notes

Figure 4. Several metabolic pathways are involved in neurodevelopment of fetal rats. (A) The combinatorial pathway of the folate cycle and the methionine metabolic pathway; (B) several compendious neurotransmitter metabolic pathways that are involved in fetal neurodevelopment; (C) general metabolic reactions of polyamines in mammalian cells; (D) the concise phosphatidylinositol metabolic pathway.

The current perception is that the application of multiple “omics” methods in biological systems will result in new biomarkers for interpreting pathophysiological variation and monitoring the effect of environmental perturbation. In other words, any single omics technique provides only one limited window into the biological activity of a biosystem. Discovering and elucidating the relationships between several variables from different biological levels is a challenging task, which has great potential for improving the way we understand bioinformation and generate biological knowledge. In this study, metabolic profiling performed using GC/MS and ICP/MS for the comprehensive study of global metabolic changes in amniotic fluid, including small-molecule metabolites and trace elements, revealed mechanistic evidence linking prenatal malnutrition to abnormal fetal neurodevelopment. Fetal neural induction, neural tube differentiation, and synaptic development appear to be affected by maternal malnutrition, as evidenced by the significantly perturbed metabolic pathways and the variation of trace elements pertaining to neurodevelopmental processes (Figure 5). In addition to studying the risk factors of fetal neurodevelopment retardation by AF metabolic profiling, we have also shown for the first time that it is possible to integrate metallomic and metabonomic profiling to interpret biological metabolic pathway changes resulting from nutritional perturbation. The combined bioinformation from metabonomic and metallomic levels provides an integrated picture of the harmful effects of prenatal malnutrition on fetal nerve and brain development with mutually supporting and validating evidence arising from each biomolecular level. Measurements made across multiple biomolecular levels as used here could potentially yield biomarker combinations that are more integrative and specific than biomolecules derived from only one omic platform. Moreover, by the cross-validation and complementation of information obtained from both metabonomic and metallomic knowledge, we enhanced our understanding of the impaired mechanism of fetal neurodevelopment resulting from malnutrition. The biological insight gained from this work may 2156

Journal of Proteome Research • Vol. 7, No. 5, 2008

Shen et al.

Figure 5. Impairment of fetal neurodevelopment hypothesis: Prenatal malnutrition may lead to abnormal neurodevelopment which involves significant alterations in endogenous metabolites and trace elements. Abbreviations used: Co, cobalt; Zn, zinc; Se, selenium; Rb, rubidium; Try, tryptophan; Gln, glutamine; Gly, glycine; T3, triiodothyronine; T4, thyroxine; DA, dopamine; 5-HT, serotonin; glu, glutamate; GABA, γ- aminobutyric acid; NMDA, N-methyl-D-aspartate receptor; PI, phosphatidylinositol.

have implications for the elucidation of a number of pathophysiological alterations in complex diseases requiring the continuous refinement of an approach to systems biology utilizing combined omic technologies. Abbreviations Used: GC/MS, gas chromatography/mass spectrometry; ICP/MS, inductively coupled plasma mass spectrometry; PCA, principal component analysis; AF, amniotic fluid; KEGG, Kyoto Encyclopedia of Gene and Genome; MVDA, multivariate data analysis.

Acknowledgment. This study was supported by the National Basic Research Program of China (Program 973, Project Number 2007CB914700) and Research Grant No. 2006DFA02700. Supporting Information Available: Table S1 shows the diet of pregnant rats during the entire gestation period. Table S2 lists microwave oven program for AF sample decomposition, and Table S3 lists the identified 98 endogenous metabolites in the rat AF sample. This material is available free of charge via the Internet at http://pubs.acs.org. References (1) Guerrini, I.; Thomson, A. D.; Gurling, H. D. The importance of alcohol misuse, malnutrition and genetic susceptibility on brain growth and plasticity. Neurosci. Biobehav. Rev. 2007, 31 (2), 212– 20. (2) Morgane, P. J.; Austin-LaFrance, R.; Bronzino, J.; Tonkiss, J.; DiazCintra, S.; Cintra, L.; Kemper, T.; Galler, J. R. Prenatal malnutrition and development of the brain. Neurosci. Biobehav. Rev. 1993, 17 (1), 91–128. (3) Knudsen, T. B.; Green, M. L. Response characteristics of the mitochondrial DNA genome in developmental health and disease. Birth Defects Res., Part C 2004, 72, 313–29. (4) Craig, A.; Sidaway, J.; Holmes, E.; Orton, T.; Jackson, D.; Rowlinson, R.; Nickson, J.; Tonge, R.; Wilson, I.; Nicholson, J. Systems toxicology: integrated genomic, proteomic and metabonomic analysis of methapyrilene induced hepatotoxicity in the rat. J. Proteome Res. 2006, 5 (7), 1586–601. (5) Nicholson, J. K.; Lindon, J. C.; Holmes, E. ‘Metabonomics’: understanding the metabolic responses of living systems to pathophysiological stimuli via multivariate statistical analysis of biological NMR spectroscopic data. Xenobiotica 1999, 29 (11), 1181–9. (6) Nicholson, J. K.; Wilson, I. D. Opinion: understanding ‘global’ systems biology: metabonomics and the continuum of metabolism. Nat. Rev. Drug Discovery 2003, 2 (8), 668–76.

technical notes

Metabonomic and Metallomic Profiling (7) Chen, M.; Su, M.; Zhao, L.; Jiang, J.; Liu, P.; Cheng, J; Lai, Y.; Liu, Y.; Jia, W. Metabonomic study of aristolochic acid-induced nephrotoxicity in rats. J. Proteome Res. 2006, 5 (4), 995–1002. (8) Wang, X.; Su, M.; Qiu, Y.; Ni, Y.; Zhao, T.; Zhou, M.; Zhao, A.; Yang, S.; Zhao, L.; Jia, W. Metabolic regulatory network alterations in response to acute cold stress and ginsenoside intervention. J. Proteome Res. 2007, 6 (9), 3449–55. (9) Young, V. R. Trace element biology: the knowledge base and its application for the nutrition of individuals and populations. J. Nutr. 2003, 133 (5 Suppl 1), 1581S–7S. (10) Szpunar, J. Metallomics: a new frontier in analytical chemistry. Anal. Bioanal. Chem. 2004, 378 (1), 54–6. (11) Szpunar, J.; Lobinski, R.; Prange, A. Hyphenated techniques for elemental speciation in biological systems. Appl. Spectrosc. 2003, 57 (3), 102A–112A. (12) Pere, M. C. Materno-foetal exchanges and utilisation of nutrients by the foetus: comparison between species. Reprod. Nutr. Dev. 2003, 43 (1), 1–15. (13) Liang, H.; Xiong, W.; Zhang, Z. Effect of maternal food restriction during gestation on early development of F1 and F2 offspring in the rat-like hamster (Cricetulus triton). Zoology (Jena) 2007, 110 (2), 118–26. (14) A, J.; Trygg, J.; Gullberg, J.; Johansson, A. I.; Jonsson, P.; Antti, H.; Marklund, S. L.; Moritz, T. Extraction and GC/MS analysis of the human blood plasma metabolome. Anal. Chem. 2005, 77 (24), 8086–94. (15) Qiu, Y.; Su, M.; Liu, Y.; Chen, M.; Gu, J.; Zhang, J.; Jia, W. Application of ethyl chloroformate derivatization for gas chromatography-mass spectrometry based metabonomic profiling. Anal. Chim. Acta 2007, 583 (2), 277–83. (16) Morita, H.; Kita, T.; Umeno, M.; Morita, M.; Yoshinaga, J.; Okamoto, K. Analysis of serum elements and the contaminations from devices used for serum preparation by inductively coupled plasma mass spectrometry. Sci. Total Environ. 1994, 151 (1), 9–17. (17) Ni, Y.; Su, M.; Qiu, Y.; Chen, M.; Liu, Y.; Zhao, A.; Jia, W. Metabolic profiling using combined GC-MS and LC-MS provides a systems understanding of aristolochic acid-induced nephrotoxicity in rat. FEBS Lett. 2007, 581 (4), 707–11. (18) Steegers-Theunissen, R. P.; Wathen, N. C.; Eskes, T. K.; van RaaijSelten, B.; Chard, T. Maternal and fetal levels of methionine and homocysteine in early human pregnancy. Br. J. Obstet. Gynaecol. 1997, 104 (1), 20–4. (19) Keating, J. N.; Wada, L.; Stokstad, E. L.; King, J. C. Folic acid: effect on zinc absorption in humans and in the rat. Am. J. Clin. Nutr. 1987, 46 (5), 835–9. (20) Almqvist, P. M.; Akesson, E.; Wahlberg, L. U.; Pschera, H.; Seiger, A.; Sundstrom, E. First trimester development of the human nigrostriatal dopamine system. Exp. Neurol. 1996, 139 (2), 227– 37.

(21) Verney, C.; Milosevic, A.; Alvarez, C.; Berger, B. Immunocytochemical evidence of well-developed dopaminergic and noradrenergic innervations in the frontal cerebral cortex of human fetuses at midgestation. J. Comp. Neurol. 1993, 336 (3), 331–44. (22) Wiggins, R. C.; Fuller, G.; Enna, S. J. Undernutrition and the development of brain neurotransmitter systems. Life Sci. 1984, 35 (21), 2085–94. (23) Sokoloff, P.; Guillin, O.; Diaz, J.; Carroll, P.; Griffon, N. Brain-derived neurotrophic factor controls dopamine D3 receptor expression: implications for neurodevelopmental psychiatric disorders. Neurotox. Res. 2002, 4 (7–8), 671–678. (24) Gambarana, C.; Ghiglieri, O.; Masi, F.; Scheggi, S.; Tagliamonte, A.; De Montis, M. G. The effects of long-term administration of rubidium or lithium on reactivity to stress and on dopamine output in the nucleus accumbens in rats. Brain Res. 1999, 826 (2), 200–9. (25) Ngo, D. B.; Dikassa, L.; Okitolonda, W.; Kashala, T. D.; Gervy, C.; Dumont, J.; Vanovervelt, N.; Contempre, B.; Diplock, A. T.; Peach, S.; Vanderpas, J. Selenium status in pregnant women of a rural population (Zaire) in relationship to iodine deficiency. Trop. Med. Int. Health 1997, 2 (6), 572–81. (26) Morreale de Escobar, G.; Obregon, M. J.; Escobar del Rey, F. Role of thyroid hormone during early brain development. Eur. J. Endocrinol. 2004, 151 (Suppl 3), U25–37. (27) Wang, F. D.; Zhao, F. J.; Jing, N. H. Effect of dietary zinc on microtubule-associated protein 2 expression in the brain of mice. Sheng Li Xue Bao 1999, 51 (5), 495–500. (28) Wallace, H. M.; Fraser, A. V.; Hughes, A. A perspective of polyamine metabolism. Biochem. J. 2003, 376 (Pt 1), 1–14. (29) Hudson, C. J.; Young, L. T.; Li, P. P.; Warsh, J. J. CNS signal transduction in the pathophysiology and pharmacotherapy of affective disorders and schizophrenia. Synapse 1993, 13 (3), 278– 93. (30) Burek, M. J.; Oppenheim, R. W. Programmed cell death in the developing nervous system. Brain Pathol. 1996, 6 (4), 427–46. (31) Wu, C.; Singh, S. K.; Dias, P.; Kumar, S.; Mann, D. M. Activated astrocytes display increased 5-HT2a receptor expression in pathological states. Exp. Neurol. 1999, 158 (2), 529–33. (32) Yan, W.; Wilson, C. C.; Haring, J. H. Effects of neonatal serotonin depletion on the development of rat dentate granule cells. Brain Res. Dev. Brain Res. 1997, 98 (2), 177–84. (33) Lujan, R.; Shigemoto, R.; Lopez-Bendito, G. Glutamate and GABA receptor signalling in the developing brain. Neuroscience 2005, 130 (3), 567–80. (34) Lauder, J. M. Hormonal and humoral influences on brain development. Psychoneuroendocrinology 1983, 8 (2), 121–55.

PR700776C

Journal of Proteome Research • Vol. 7, No. 5, 2008 2157