Article pubs.acs.org/jpr
Plasma Metabolite Profiles of Alzheimer’s Disease and Mild Cognitive Impairment Gang Wang,†,# Yi Zhou,†,# Feng-Jie Huang,‡ Hui-Dong Tang,† Xu-Hua Xu,† Jia-Jian Liu,‡ Ying Wang,† Yu-Lei Deng,† Ru-Jing Ren,† Wei Xu,† Jian-Fang Ma,† Yi-Nan Zhang,‡ Ai-Hua Zhao,‡ Sheng-Di Chen,*,†,§ and Wei Jia*,‡ †
Department of Neurology and Institute of Neurology, Rui Jin Hospital affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai 200025, China ‡ Center for Translational Medicine, Shanghai Jiao Tong University Affiliated Sixth People’s Hospital, Shanghai 200233, China § Key Laboratory of Stem Cell Biology and Lab of Neurodegenerative Diseases, Institute of Health Science, Shanghai Institutes of Biological Sciences, Chinese Academy of Science and Shanghai Jiao Tong University School of Medicine, Shanghai 200025, China S Supporting Information *
ABSTRACT: Previous studies have demonstrated altered metabolites in samples of Alzheimer’s disease (AD) patients. However, the sample size from many of them is relatively small and the metabolites are relatively limited. Here we applied a comprehensive platform using ultraperformance liquid chromatography-time-of-flight mass spectrometry and gas chromatography-time-of-flight mass spectrometry to analyze plasma samples from AD patients, amnestic mild cognitive impairment (aMCI) patients, and normal controls. A biomarker panel consisting of six plasma metabolites (arachidonic acid, N,Ndimethylglycine, thymine, glutamine, glutamic acid, and cytidine) was identified to discriminate AD patients from normal control. Another panel of five plasma metabolites (thymine, arachidonic acid, 2-aminoadipic acid, N,N-dimethylglycine, and 5,8-tetradecadienoic acid) was able to differentiate aMCI patients from control subjects. Both biomarker panels had good agreements with clinical diagnosis. The 2 panels of metabolite markers were all involved in fatty acid metabolism, one-carbon metabolism, amino acid metabolism, and nucleic acid metabolism. Additionally, no altered metabolites were found among the patients at different stages, as well as among those on anticholinesterase medication and those without anticholinesterase medication. These findings provide a comprehensive global plasma metabolite profiling and may contribute to making early diagnosis as well as understanding the pathogenic mechanism of AD and aMCI. KEYWORDS: Alzheimer’s disease, amnestic mild cognitive impairment, metabolomics, plasma, biomarkers
■
unique chemical fingerprints that cellular processes leave behind.7,8 Recently, metabolomics has become a powerful profiling approach with the aim to discover novel biomarkers and to understand the mechanistic pathways underlying a pathophysiological state. The common analytical platforms used in metabolomics include nuclear magnetic resonance (NMR) spectroscopy and mass spectrometry (MS) often coupled with gas chromatography (GC) and/or liquid chromatography (LC). Among these platforms, GC-MS specializes in measurement of volatile compounds such as fatty acids, whereas LC with a relatively wider universality is now used predominantly in metabolomics.9 Today, a series of studies have demonstrated altered lipids at the different stages of AD in brain tissue samples and plasma samples.10−12 Other research also found that amino acids,
INTRODUCTION The pathogenesis of Alzheimer’s disease (AD) has been associated with multiple pathways, including altered amyloid precursor protein (APP) metabolism, tau hyperphosphorylation, neuroinflammation, mitochondrial dysfunction, and oxidative stress.1−3 Amnestic mild cognitive impairment (aMCI) is believed to be an intermediate state between normal cognition and AD.4 In recent years, multiple studies have been conducted using neuroimaging, genetic testing, and neurochemical testing of body fluids with reasonable success.5,6 However, biomarker discovery for the accurate diagnosis of aMCI and AD remains to be one of the urgent research initiatives. Metabolomics is an emerging profiling method with unbiased identification and state-specific quantification of low molecular weight molecules (less than 1 Kda) in a given biological compartment. This technology plays an increasingly important role in biological research because of the ability of revealing the © 2014 American Chemical Society
Received: January 28, 2014 Published: April 2, 2014 2649
dx.doi.org/10.1021/pr5000895 | J. Proteome Res. 2014, 13, 2649−2658
Journal of Proteome Research
Article
Blood Collection and Pretreatment
glycoproteins, and choline metabolites were significantly altered.13−17 These studies exemplified metabolomic applications in the studies of AD and other cognitive disorders, providing the meaningful cues for pathogenetic mechanisms. However, the sample size from many of the previous studies is relatively small, and the metabolites are relatively limited.18−22 To date, there has not been a comprehensive global profiling of small molecule metabolites in plasma using LC-MS and GCMS in the study of AD and aMCI. Here we analyzed the plasma metabolites in AD patients, aMCI patients, and normal controls, aiming to identify a distinct metabolomics profile and specific plasma biomarkers for the early diagnosis of AD and aMCI.
■
Fasting venous blood was collected from all the participants in ethylenediaminetetraacetic acid (EDTA) anticoagulant tubes and then centrifuged at 8000g for 10 min. The obtained plasma was frozen at −80 °C until utilized for the LC/MS and GC/MS analysis. Pooled quality control (QC) samples were prepared by mixing 20 μL of each the plasma sample. APOEε4 Genotyping
Genomic DNA was extracted from peripheral whole blood cells using the standardized phenol/chloroform extraction method. Phenotyping of APOEε4 was performed by PCR-restriction fragment length polymorphism assays (PCR-RFLP) as described in the research of Zivelin et al.25
MATERIALS AND METHODS
Plasma Sample Preparation and Data Analysis by GC-TOFMS
Subjects
Following our previous similar procedure,26 a 100 μL aliquot of plasma sample was spiked with two internal standards (10 μL of L-2-chlorophenylalanine in water, 0.3 mg/mL; 10 μL of heptadecanoic acid in methanol, 1 mg/mL) and vortexed for 10 s. The mixed solution was extracted with 300 μL of methanol/ chloroform (3:1) and vortexed for 30 s. After the samples were stored for 10 min at −20 °C, they were centrifuged at 8000 rpm for 10 min. An aliquot of the 300 μL supernatant was transferred to a glass sampling vial to vacuum-dry at room temperature. The residue was subjected to a derivatization procedure with 80 μL of methoxyamine (15 mg/mL in pyridine) for 90 min at 30 °C, followed by 80 μL of N,Obis(trimethylsilyl)trifluoroacetamide (BSTFA) (1% trimethylchlorosilane (TMCS)) for 60 min at 70 °C. When the reaction was finished, the samples were placed at room temperature for 1 h waiting for GC-TOFMS analysis. The samples were analyzed by the GC-TOFMS (Pegasus HT, Leco Corp., St. Joseph, MI; electron ionization (EI) mode) in the order “AD-aMCI-control”. One QC sample and one blank vial were run after each 10 plasma samples. The injection volume was 1 μL with a splitless mode. A DB-5 ms capillary column (30 m × 250 μm i.d., 0.25-μm film thickness; 5% diphenyl cross-linked 95% dimethylpolysiloxane) was used for the separation of metabolites. The carrier gas was helium (99.9996%) with a constant flow rate of 1 mL/min. The GC oven temperature was started at 80 °C for 2 min, and then the temperature was increased to 180 °C at 10 °C/min, to 230 °C at 6 °C/min, and lastly to 295 °C at 40 °C/min, which was maintained for 8 min. The temperature of the injection, the transfer interface, and the ion source were set to 270, 260, and 220 °C, respectively. The mass range was set to 30−600 with electron impact ionization (70 eV), and the acquisition rate was 20 spectrum/second. The obtained files from GC-TOFMS analysis were exported in NetCDF format by ChromaTOF software (v4.44, Leco Co., Los Angeles, CA). CDF files were pretreated through baseline correction, denoising, smoothing, alignment, time-window splitting, and multivariate curve resolution based on the kit developed by MATLAB 7.0 (The MathWorks Inc. Natick, MA), R 10.2 (Lucent Technologies) and JavaSE 1.6 (Sun Microsystems). The internal standard and QC were used for data quality control (reproducibility), and QC was used for data normalization. The ion peaks generated by the internal standard were also removed.
A total of 57 patients with AD and 58 patients with aMCI were recruited from the Department of Neurology, Rui Jin Hospital affiliated to Shanghai Jiao Tong University School of Medicine. Additionally, 57 community-dwelling elderly individuals (healthy volunteers) of comparable age and sex were included as controls. No concomitant medication intended to influence plasma lipid levels was permitted. The study was approved by the Research Ethics Committee, Rui Jin Hospital affiliated to Shanghai Jiao Tong University School of Medicine, China. Written informed consent was obtained from each participant. It took 1 year for the overall sampling collection. All AD patients were diagnosed as probable AD according to NINCDS-ADRDA (National Institute for Neurological Diseases and Stroke/Alzheimer’s Disease and Related Disorders Association) criteria.23 Diagnosis was made on the basis of clinical and neuropsychological examination. Cognitive impairment (CI) of the patients was progressive and present in two or more areas of cognition. The deficits were ascertained by interview of the caregiver, dated back to when the cognitive symptoms were first noted. Brain magnetic resonance imaging scanning, screening for hypertension and depression, and blood tests were also performed to exclude other diseases capable of producing a dementia syndrome. The severity of the cognitive decline was graded according to the mini-mental state examination (MMSE) and the AD patients were categorized as mild, moderate or severe. Of the 57 AD subjects, 24 had mild AD (MMSE ⩾ 20), 26 had moderate AD (MMSE ⩾ 10, MMSE < 20), and 7 had severe AD (MMSE < 10). The aMCI patients were diagnosed following Peterson clinical criteria originally proposed by Petersen and coworkers.24 All the subjects had complaints of memory loss either by themselves or by their family members, but no impairment of daily activities and no dementia were reported. Neuropsychological tests of memory showed objective impairment (≥1.5 SD below the age-appropriate mean) and without significant functional decline. MMSE and clinical dementia rating (CDR) score were selected to describe cognitive status of aMCI subjects. All the tested subjects reached 0.5 in CDR score. The same as AD subjects, aMCI patients had imaging and laboratory tests to confirm the absence of other possible pathologies underlying the symptoms. The control subjects were evaluated by a neurologist to confirm that there was no history of dementia or other neurologic diseases. 2650
dx.doi.org/10.1021/pr5000895 | J. Proteome Res. 2014, 13, 2649−2658
Journal of Proteome Research
Article
Table 1. Subject Characteristicsa subjects n age (mean years ± SD) male/female (n/n) education (mean years ± SD) MMSE (mean score ± SD)
cholinesterase inhibitor user n(%) Apo ε 4 (+)-carriers n(%)
AD
aMCI
control
57 74.16 ± 9.05 24/33 10.45 ± 4.92 17.88 ± 5.55 AD vs aMCI
Plasma Metabolite Profiles of Alzheimer’s Disease and Mild Cognitive Impairment Gang Wang,†,# Yi Zhou,†,# Feng-Jie Huang,‡ Hui-Dong Tang,† Xu-Hua Xu,† Jia-Jian Liu,‡ Ying Wang,† Yu-Lei Deng,† Ru-Jing Ren,† Wei Xu,† Jian-Fang Ma,† Yi-Nan Zhang,‡ Ai-Hua Zhao,‡ Sheng-Di Chen,*,†,§ and Wei Jia*,‡ †
Department of Neurology and Institute of Neurology, Rui Jin Hospital affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai 200025, China ‡ Center for Translational Medicine, Shanghai Jiao Tong University Affiliated Sixth People’s Hospital, Shanghai 200233, China § Key Laboratory of Stem Cell Biology and Lab of Neurodegenerative Diseases, Institute of Health Science, Shanghai Institutes of Biological Sciences, Chinese Academy of Science and Shanghai Jiao Tong University School of Medicine, Shanghai 200025, China S Supporting Information *
ABSTRACT: Previous studies have demonstrated altered metabolites in samples of Alzheimer’s disease (AD) patients. However, the sample size from many of them is relatively small and the metabolites are relatively limited. Here we applied a comprehensive platform using ultraperformance liquid chromatography-time-of-flight mass spectrometry and gas chromatography-time-of-flight mass spectrometry to analyze plasma samples from AD patients, amnestic mild cognitive impairment (aMCI) patients, and normal controls. A biomarker panel consisting of six plasma metabolites (arachidonic acid, N,Ndimethylglycine, thymine, glutamine, glutamic acid, and cytidine) was identified to discriminate AD patients from normal control. Another panel of five plasma metabolites (thymine, arachidonic acid, 2-aminoadipic acid, N,N-dimethylglycine, and 5,8-tetradecadienoic acid) was able to differentiate aMCI patients from control subjects. Both biomarker panels had good agreements with clinical diagnosis. The 2 panels of metabolite markers were all involved in fatty acid metabolism, one-carbon metabolism, amino acid metabolism, and nucleic acid metabolism. Additionally, no altered metabolites were found among the patients at different stages, as well as among those on anticholinesterase medication and those without anticholinesterase medication. These findings provide a comprehensive global plasma metabolite profiling and may contribute to making early diagnosis as well as understanding the pathogenic mechanism of AD and aMCI. KEYWORDS: Alzheimer’s disease, amnestic mild cognitive impairment, metabolomics, plasma, biomarkers
■
unique chemical fingerprints that cellular processes leave behind.7,8 Recently, metabolomics has become a powerful profiling approach with the aim to discover novel biomarkers and to understand the mechanistic pathways underlying a pathophysiological state. The common analytical platforms used in metabolomics include nuclear magnetic resonance (NMR) spectroscopy and mass spectrometry (MS) often coupled with gas chromatography (GC) and/or liquid chromatography (LC). Among these platforms, GC-MS specializes in measurement of volatile compounds such as fatty acids, whereas LC with a relatively wider universality is now used predominantly in metabolomics.9 Today, a series of studies have demonstrated altered lipids at the different stages of AD in brain tissue samples and plasma samples.10−12 Other research also found that amino acids,
INTRODUCTION The pathogenesis of Alzheimer’s disease (AD) has been associated with multiple pathways, including altered amyloid precursor protein (APP) metabolism, tau hyperphosphorylation, neuroinflammation, mitochondrial dysfunction, and oxidative stress.1−3 Amnestic mild cognitive impairment (aMCI) is believed to be an intermediate state between normal cognition and AD.4 In recent years, multiple studies have been conducted using neuroimaging, genetic testing, and neurochemical testing of body fluids with reasonable success.5,6 However, biomarker discovery for the accurate diagnosis of aMCI and AD remains to be one of the urgent research initiatives. Metabolomics is an emerging profiling method with unbiased identification and state-specific quantification of low molecular weight molecules (less than 1 Kda) in a given biological compartment. This technology plays an increasingly important role in biological research because of the ability of revealing the © 2014 American Chemical Society
Received: January 28, 2014 Published: April 2, 2014 2649
dx.doi.org/10.1021/pr5000895 | J. Proteome Res. 2014, 13, 2649−2658
Journal of Proteome Research
Article
Blood Collection and Pretreatment
glycoproteins, and choline metabolites were significantly altered.13−17 These studies exemplified metabolomic applications in the studies of AD and other cognitive disorders, providing the meaningful cues for pathogenetic mechanisms. However, the sample size from many of the previous studies is relatively small, and the metabolites are relatively limited.18−22 To date, there has not been a comprehensive global profiling of small molecule metabolites in plasma using LC-MS and GCMS in the study of AD and aMCI. Here we analyzed the plasma metabolites in AD patients, aMCI patients, and normal controls, aiming to identify a distinct metabolomics profile and specific plasma biomarkers for the early diagnosis of AD and aMCI.
■
Fasting venous blood was collected from all the participants in ethylenediaminetetraacetic acid (EDTA) anticoagulant tubes and then centrifuged at 8000g for 10 min. The obtained plasma was frozen at −80 °C until utilized for the LC/MS and GC/MS analysis. Pooled quality control (QC) samples were prepared by mixing 20 μL of each the plasma sample. APOEε4 Genotyping
Genomic DNA was extracted from peripheral whole blood cells using the standardized phenol/chloroform extraction method. Phenotyping of APOEε4 was performed by PCR-restriction fragment length polymorphism assays (PCR-RFLP) as described in the research of Zivelin et al.25
MATERIALS AND METHODS
Plasma Sample Preparation and Data Analysis by GC-TOFMS
Subjects
Following our previous similar procedure,26 a 100 μL aliquot of plasma sample was spiked with two internal standards (10 μL of L-2-chlorophenylalanine in water, 0.3 mg/mL; 10 μL of heptadecanoic acid in methanol, 1 mg/mL) and vortexed for 10 s. The mixed solution was extracted with 300 μL of methanol/ chloroform (3:1) and vortexed for 30 s. After the samples were stored for 10 min at −20 °C, they were centrifuged at 8000 rpm for 10 min. An aliquot of the 300 μL supernatant was transferred to a glass sampling vial to vacuum-dry at room temperature. The residue was subjected to a derivatization procedure with 80 μL of methoxyamine (15 mg/mL in pyridine) for 90 min at 30 °C, followed by 80 μL of N,Obis(trimethylsilyl)trifluoroacetamide (BSTFA) (1% trimethylchlorosilane (TMCS)) for 60 min at 70 °C. When the reaction was finished, the samples were placed at room temperature for 1 h waiting for GC-TOFMS analysis. The samples were analyzed by the GC-TOFMS (Pegasus HT, Leco Corp., St. Joseph, MI; electron ionization (EI) mode) in the order “AD-aMCI-control”. One QC sample and one blank vial were run after each 10 plasma samples. The injection volume was 1 μL with a splitless mode. A DB-5 ms capillary column (30 m × 250 μm i.d., 0.25-μm film thickness; 5% diphenyl cross-linked 95% dimethylpolysiloxane) was used for the separation of metabolites. The carrier gas was helium (99.9996%) with a constant flow rate of 1 mL/min. The GC oven temperature was started at 80 °C for 2 min, and then the temperature was increased to 180 °C at 10 °C/min, to 230 °C at 6 °C/min, and lastly to 295 °C at 40 °C/min, which was maintained for 8 min. The temperature of the injection, the transfer interface, and the ion source were set to 270, 260, and 220 °C, respectively. The mass range was set to 30−600 with electron impact ionization (70 eV), and the acquisition rate was 20 spectrum/second. The obtained files from GC-TOFMS analysis were exported in NetCDF format by ChromaTOF software (v4.44, Leco Co., Los Angeles, CA). CDF files were pretreated through baseline correction, denoising, smoothing, alignment, time-window splitting, and multivariate curve resolution based on the kit developed by MATLAB 7.0 (The MathWorks Inc. Natick, MA), R 10.2 (Lucent Technologies) and JavaSE 1.6 (Sun Microsystems). The internal standard and QC were used for data quality control (reproducibility), and QC was used for data normalization. The ion peaks generated by the internal standard were also removed.
A total of 57 patients with AD and 58 patients with aMCI were recruited from the Department of Neurology, Rui Jin Hospital affiliated to Shanghai Jiao Tong University School of Medicine. Additionally, 57 community-dwelling elderly individuals (healthy volunteers) of comparable age and sex were included as controls. No concomitant medication intended to influence plasma lipid levels was permitted. The study was approved by the Research Ethics Committee, Rui Jin Hospital affiliated to Shanghai Jiao Tong University School of Medicine, China. Written informed consent was obtained from each participant. It took 1 year for the overall sampling collection. All AD patients were diagnosed as probable AD according to NINCDS-ADRDA (National Institute for Neurological Diseases and Stroke/Alzheimer’s Disease and Related Disorders Association) criteria.23 Diagnosis was made on the basis of clinical and neuropsychological examination. Cognitive impairment (CI) of the patients was progressive and present in two or more areas of cognition. The deficits were ascertained by interview of the caregiver, dated back to when the cognitive symptoms were first noted. Brain magnetic resonance imaging scanning, screening for hypertension and depression, and blood tests were also performed to exclude other diseases capable of producing a dementia syndrome. The severity of the cognitive decline was graded according to the mini-mental state examination (MMSE) and the AD patients were categorized as mild, moderate or severe. Of the 57 AD subjects, 24 had mild AD (MMSE ⩾ 20), 26 had moderate AD (MMSE ⩾ 10, MMSE < 20), and 7 had severe AD (MMSE < 10). The aMCI patients were diagnosed following Peterson clinical criteria originally proposed by Petersen and coworkers.24 All the subjects had complaints of memory loss either by themselves or by their family members, but no impairment of daily activities and no dementia were reported. Neuropsychological tests of memory showed objective impairment (≥1.5 SD below the age-appropriate mean) and without significant functional decline. MMSE and clinical dementia rating (CDR) score were selected to describe cognitive status of aMCI subjects. All the tested subjects reached 0.5 in CDR score. The same as AD subjects, aMCI patients had imaging and laboratory tests to confirm the absence of other possible pathologies underlying the symptoms. The control subjects were evaluated by a neurologist to confirm that there was no history of dementia or other neurologic diseases. 2650
dx.doi.org/10.1021/pr5000895 | J. Proteome Res. 2014, 13, 2649−2658
Journal of Proteome Research
Article
Table 1. Subject Characteristicsa subjects n age (mean years ± SD) male/female (n/n) education (mean years ± SD) MMSE (mean score ± SD)
cholinesterase inhibitor user n(%) Apo ε 4 (+)-carriers n(%)
AD
aMCI
control
57 74.16 ± 9.05 24/33 10.45 ± 4.92 17.88 ± 5.55 AD vs aMCI
