High-Throughput Metabolomics for Discovering Potential Metabolite

Jul 28, 2017 - Fax: +86-451-82110818. .... After 30 min, the mice were placed back in the same starting arm for the ... The peak picking parameter's r...
2 downloads 0 Views 1MB Size
Subscriber access provided by UNIV OF NEWCASTLE

Article

High-throughput metabolomics for discovering potential metabolite biomarkers and metabolic mechanism from APPswe/PS1dE9 transgenic model of Alzheimer's disease Jingbo Yu, Ling Kong, Aihua Zhang, Ying Han, Zhidong Liu, Hui Sun, Liang Liu, and Xijun Wang J. Proteome Res., Just Accepted Manuscript • DOI: 10.1021/acs.jproteome.7b00206 • Publication Date (Web): 28 Jul 2017 Downloaded from http://pubs.acs.org on July 30, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Proteome Research is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

High-throughput metabolomics for discovering potential metabolite biomarkers and metabolic mechanism from APPswe/PS1dE9 transgenic model of Alzheimer's disease

Jingbo Yu1, Ling Kong1, Aihua Zhang1, Ying Han1, Zhidong Liu1, Hui Sun1, Liang Liu2, Xijun Wang1,2* 1. Sino-America Chinmedomics Technology Collaboration Center, National TCM Key Laboratory of Serum Pharmacochemistry, Chinmedomics Research Center of State Administration of TCM, Metabolomics Laboratory, Department of Pharmaceutical Analysis, Heilongjiang University of Chinese Medicine, Heping Road 24, Harbin 150040; China. 2. State Key Laboratory of Quality Research in Chinese Medicine, Macau University of Science and Technology, AvenidaWai Long, Taipa, Macau

Correspondence Prof. Xijun Wang Sino-America Chinmedomics Technology Collaboration Center, National TCM Key Laboratory of Serum Pharmacochemistry, Chinmedomics Research Center of State Administration of TCM, Metabolomics Laboratory, Department of Pharmaceutical Analysis, Heilongjiang University of Chinese Medicine, Heping Road 24, Harbin 150040; State Key Laboratory of Quality Research in Chinese Medicine, Macau University of Science and Technology, AvenidaWai Long, Taipa, Macau. Email: [email protected]; Tel. & Fax +86-451-82110818

1 ACS Paragon Plus Environment

Journal of Proteome Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 23

ABSTRACT Alzheimer’s disease (AD), a neurodegenerative disorder, is the major form of dementia. As AD is an irreversible disease, it is necessary to reinforce earlier intervention. However, the potential biomarkers on preclinical AD are still not clearly. In this study, urinary metabolomics based on ultra-high performance liquid chromatography coupled with quadruple time-of-flight mass spectrometry was performed for delineating the metabolic changes and potential early biomarkers in APPswe/PS1dE9 (APP/PS1) transgenic mice. Compared to wide-type, a total of 24 differential metabolites were identified in transgenic mice using multivariate statistical analysis. Among them, 10 metabolites were significantly up-regulated and 14 metabolites were down-regulated. Based on these potential biomarkers, metabolic pathway analysis found that pentose and glucuronate interconversions, glyoxylate and dicarboxylate metabolism, starch and sucrose metabolism, citrate cycle, tryptophan metabolism and arginine and proline metabolism were disturbed in APP/PS1 mice. Our study manifested that endogenous metabolites in the urine of APP/PS1 mice have changed priors to the emergence of learning and cognitive impairment, which may be associated with abnormal NO production pathways and metabolic disorders of the monoaminergic neurotransmitters. In conclusion, this study showed that metabolomics provides an early indicator of the disease occurrence. Keywords: Alzheimer’s disease; transgenic APPswe/PS1dE9; metabolomics; mass spectrometry; biomarkers

2 ACS Paragon Plus Environment

Page 3 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

Introduction Alzheimer’s disease (AD), an irreversible neurodegenerative disorder involving cognitive impairment, accounts for 60-80% of all dementia cases.1 It is characterized by neuropathological hallmarks including extracellular accumulation of amyloid-β (Aβ) peptide, a deposit of tau protein in intracellular and neuronal, axonal and synaptic loss and dysfunction.2-4 The National Institute on Aging and the Alzheimer’s association developed recommendations to define the preclinical stages of AD and hoped to determine the factors which best predict the risk of AD progression.5 As is known to all, the pathophysiology of AD, such as large amounts of synaptic loss, cognitive decline, and behavioral changes is related to Aβ accumulation and oligomericspecies.6,7 Early-onset familial AD, an autosomal dominant AD, is associated with mutation in genes presenilins 1 and 2 (PS1 and PS2) and amyloid precursor protein (APP), which results in the subsequent accumulation of Aβ.8 APPswe/PS1dE9 (APP/PS1) transgenic mouse model, co-expressing mutated PS1 and APP, have been widely used in studies to unveil the early pathological mechanisms and assess the efficacy of AD therapies.9 Since the underlying pathogenesis mechanism of AD is complicated and it is irreversible after onset, a convenient and accurate method for diagnosis of AD in the early stages is urgently needed. Metabolomics is used to identify metabolites in whole organisms, relating gene expression to phenotypic outcome.10 Biomarkers are dependable predictors of a disease process. It is significant to identify an effective biomarker based on APP/PS1 gene for early detection of AD. Numerous studies have been performed to examine changes of metabolites in transgenic AD mice. Metabolomics analysis of serum samples from APP/PS1 mice has found that a number of metabolites were perturbed, which could be associated with disturbed amino acids metabolism, energy-related failures, abnormal homeostasis of membrane phospholipids and abnormal metabolism of cholesterol.11 A NMR-based metabolomics study found that increases in 3-hydroxykynurenine, homogentisate and allantoin in Tg2576 transgenic mice urine in 4 months, which highlights the correlation between oxidative stress and AD.12 In this study, urine samples from APP/PS1 mice (n = 30) and wild-type controls (n = 30) in 2-month old were analyzed using a metabolomics technology based on ultra-high performance liquid chromatography-mass/mass spectrometry. Then, all spectra were processed using multivariable analysis, including Principal Component Analysis (PCA) and Orthogonal Partial Least Square-Discriminant Analysis (OPLS-DA) algorithms, to determine the changes in urine metabolites in holistic perspective. The results clearly showed that the urine metabolic profiles were disturbed by APP/PS1 genes as early as 2 months priors to any behavioral changes appear and 24 differential metabolites were identified. These results may have implications for earlier AD diagnosis and intervention of AD.

MATERIALS AND METHODS Animals Male mice were purchased from the National Resource Center of Mutant Mouse (NRCMM, China), 3 ACS Paragon Plus Environment

Journal of Proteome Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 23

including double transgenic mouse B6C3-Tg (APPswe, PSEN1dE9) 85Dbo /JNju (J004462) and wild-type mice (J004462). All experimental animals were bred in an Association for Assessment and Accreditation of Laboratory Animal Care-approved, temperature (24 ± 2 ℃), humidity (60 ± 5 %), and 12 h dark/light cycles controlled facility with ad libitum access to food and water. For this study, APP/PS1 mice at the age of 2-month old, and aged matched WT littermate controls (n = 30) were used. The study was approved by the Ethical Committee of Heilongjiang University of Chinese Medicine and was conducted according to the principles expressed in the Declaration of Helsinki. Behavioral assessment Modified Y-maze test A modified Y-maze test was performed as previously described to test spatial working memory in mice.13 Briefly, the apparatus consisted of three arms (40 × 8 × 18 cm): the start arm, the novel arm and the other arm. A test phase trial was performed 30 min after the sample phase trial (Fig. 1A). During the training trial (10 min), the mice were place into the start arm to explore maze, meanwhile the novel arm was blocked. After 30 min, the mice were placed back in the same starting arm for the test trial (10 min), with free access to all three arms. Mouse was randomly placed in different arms as starting points. The apparatus was cleaned using 75% ethanol between trials to prevent from odor interference. Performance of mice was tracked and recorded by video for later analysis. Percent time spent in the novel arm was calculated using Ethovision XT 11.5 (Noldus, Wageningen, Netherlands). Morris water maze test Spatial learning and memory of mice was measured using Morris water maze test as previously described.14 The system consisted of a circular water tank (120 cm diameter and 50 cm high) and the water temperature was maintained at (22 ± 1℃). The test comprised place navigation test and probe trial. For the place navigation test, a platform (10 cm diameter) was located in the Northeast quadrant and submerged 1-2 cm below the opaque water surface. Mice were trained four times per day for 5 consecutive days. In each trial, they were given 60 s to find a hidden platform and the trail was terminated when the mice climbed onto the platform or after 60 s. Each mouse was allowed to stay on the platform for 20 s. On the sixth day, the probe trial was conducted with the platform removed. The mice were allowed to swim freely for 60 s. The time that was taken for a mouse in the target quadrant were recorded and analyzed using Ethovision XT 11.5 (Noldus, Wageningen, Netherlands). Urine collection and preparation All animals were allowed to acclimatize in metabolism cages for 3 days prior to experiment. 12 hours overnight urine was collected at 8:00 am on age of 8 week. The samples were centrifuged at 13,000 rpm for 10 min at 4 ℃. The supernatant was diluted 1:7 with HPLC grade water, vortexed, screened with 0.22 µm filter membrane for UPLC-Q-TOF-HDMS analysis. Metabolomics analysis 4 ACS Paragon Plus Environment

Page 5 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

Chromatographic analysis was performed in Waters AcquityTM Ultra Performance LC system (Waters, Milford, USA) and equipped with quaternary pump, vacuum degasser, autosampler, diodearray detector. An aliquot of 4 µL of sample solution for ESI+ and was injected into HSS T3 1.8 µM 2.1 × 100 mm column held at 40 ℃ and the flow rate was 0.4 mL/min. The gradient mobile phase comprised of 100% acetonitrile containing 0.1% formic acid (phase A) and water with 0.1% formic acid (phase B). The following gradient was applied: 0-4 min, 1-20% A; 4-7.5 min, 20-40%; 7.5-8.5 min, 40-99%; 8.5-10.5 min, 99% A isocratic; 10.5-11min, 99-1%; 11-15min 1% A isocratic. The Mass spectrometry system was quadrupole-time-of-flight mass spectrometer (SYNAPT G2-Si HDMS, Waters Corporation, UK) equipped with an ESI ion source that operates in positive ionization mode (ESI+) and negative ionization mode (ESI-) at 50-1000 m/z in the full scan mode. The ES+ parameters are as follows: capillary voltage 3000 V, sampling cone voltage 40 V, source temperature 110 ℃, desolvation temperature 350 ℃, cone gas flow 50 L/h, and desolvation gas flow 600 L/h. The ES- parameter are as follows: capillary voltage 2200 V, sampling cone voltage 40 V, source temperature 110 ℃, desolvation temperature 350 ℃, cone gas flow 50 L/h, and desolvation gas flow 600 L/h. The external reference (leucine enkaphalin) was used to real-time correction at a concentration of 0.2 ng/mL under a flow rate of 10 µL/min via a lockspray interface, for positive mode [M+H]+ = 556.2771 and negative mode [M-H]- = 554.2615. All data were acquired using MassLynxTM (V4.1) software in centroid format. Multivariate statistical analysis and metabolite identification All mass data acquired were imported to Progenesis QI (Nonlinear Dynamics, Newcastle, U.K.) for peak alignment (RT and m/z values), peak picking and normalization. The peak picking parameters retention time limits was set from 0 to 8.5 min and normalization method was set to use total ion intensity. Then the resultant data matrices were introduced to Ezinfo 2.0 software for multivariate statistical analysis, including principal components (PCA) and orthogonal partial least-squared discriminant analysis (OPLS-DA). The VIP-plot was established from OPLS-DA to search for the difference variables which had significant contributions to classification from the datasets. Combined with T-test, we selected the variables with VIP > 1 and P < 0.05 as potential biomarkers for further identification. The molecular formula was calculated using accurate mass and MS/MS fragments information was compared with the Chemspider (http:// www.chemspider.com/), the Human Metabolome Database (HMDB) (http://www.hmdb.ca/) and METLIN (https;//metlin.scripps.edu/). The measurement error of precise molecular mass is less than 5 ppm. Metabolic pathway analysis Metabolic pathway analysis (MetPA) (http://www.metaboanalyst.ca/MetaboAnalyst/faces/Secure/upload/PathUploadView.xhtml) is a web-based analysis tool used to construct and visualize the affected metabolic pathways based on database source, including KEGG (http://www.genome.jp/kegg/) and HMDB. A metabolic network was also performed with the Ingenuity Pathway Analysis (IPA) software (Ingenuity, Redwood City, CA, USA) to explore the interplays among the potential biomarkers. Statistical analysis 5 ACS Paragon Plus Environment

Journal of Proteome Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 23

The statistical analysis was done using SPPS 17 software (IBM, NY, USA). All results are expressed as mean ± SD. Y maze and metabolomics were analyzed by student’s t-test. Morris water maze was analyzed by repeated measure ANOVA. Statistical significant values were set to P 0.05). This indicates that no impaired spatial working memory in 2-month-old APP/PS1 Tg mice. In addition, the spatial learning and memory function of all mice were assessed by Morris water maze. The results of the time required to find the hidden platform (escape latency) for all mice during the water maze acquisition training are shown in Fig. 1C. Repeated-measures ANOVA revealed no significant main effect of genotype on the escape latency over all training days in 2-month-old mice (F(1,1) = 0.075, P > 0.05), indicating that 2-month-old APP/PS1 mice had no impaired spatial learning function compared with WT mice. Spatial memory of all mice was evaluated by the probe trial performed at 24 h after the last training session. As shown in Fig1. D, APP/PS1 and WT mice got similar percent time spent in the target quadrant (P > 0.05), indicating no impaired spatial memory retention in these transgenic mice at the age of 2 months. Metabolome data processing All the urine samples were analyzed by UPLC-Q-TOF-MS using the optimal conditions described above. In the sequence table, a quality control (QC) sample was run per ten samples. Raw data were preprocessed by Progenesis QI. After peak alignment, peak picking and normalization, a total of 6182 ions in the positive mode and 5408 ions in the negative mode were detected. Then, the ions were imported into EZinfo 2.0 software for multivariate data analysis (Fig. 2). To assess the data quality of the urine samples, coefficients of variance was calculated for QC samples per metabolite feature. The coefficient of variance of 5192 metabolite features and 4087 metabolite features are less than 30% in positive ion mode and negative ion mode, respectively. Metabolic profiling analysis Representative positive and negative based peak intensity (BPI) chromatograms of urine obtained from APP/PS1 mice and WT mice are shown in Fig. 3C and Fig. 3D, respectively. The urine profiles of the BPI chromatograms of the two groups were generally similar, but several peaks were modulated. Multivariate statistical analysis was obtained to better observe the differences among these complex matrix. Multivariate statistical analysis The 3D-PCA score plot showed a trend of separation between the WT mice and APP/PS1 mice (Fig. 3A and Fig. 3B). The result suggested that metabolite perturbation of APP/PS1 mice was generated as a result of APP/PS1 genes in 2 6 ACS Paragon Plus Environment

Page 7 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

months old. In order to further distinguish between the APP/PS1 mice and WT mice, the supervised OPLS-DA was employed to divide samples into two blocks. In our study, the OPLS-DA score plot analysis separated the urine samples of APP/PS1 mice and WT mice completely (Fig. 4A and 4B) based on the difference in their metabolic profiles. Moreover, the VIP plot was performed to mine the significant variables according to a VIP value (Fig. 4C and 4D). If the VIP value of the variable is greater than 1, it means that the variable could contribute significantly to classification. Finally, 109 ions in the positive and 37 ions in the negative were selected (VIP > 1 and P < 0.05) and identified as being accountable for the discrimination between APP/PS1 mice and WT mice. Identification of candidate metabolites The candidate metabolites were tentatively identified using accurate mass and MS/MS fragments data provided by Q-TOF platform. Then, the MS and MS/MS fragments information was applied to identify the suspected metabolites combined with consulting online database, including the HMDB, METLIN, Chemspider and KEGG. Taking an example, ESI positive ion mode gave an [M + H]+ ion at m/z 340.1014, indicating the formula C15H18NO8. At its MS/MS data, it produced fragment ions at m/z 322[M + H - H2O]+, 176[M + H - C9H10NO2]+, 164[M + H - C6H8O6]+, 146[M + H - C6H10O7]+, 132[M + H - C7H12O7]+ (Fig. 5.), respectively. According to the protocol described above, 24 potential biomarkers were tentatively identified and characterized, and these were summarized in Table S1 and Table S2. Compared with WT mice, ten metabolites were up regulated, including spermic acid, 2,4-guanidinobutanoic acid, nicotinuric acid, L-isoleucyl-L-proline, L-2,3-dihydrodipicolinate, 3,4-dihydroxy-phenylglycol o-sulfate, N-acetyl-Ltyrosine, 5-hydroxyindoleacetic acid, 3-methoxy-benzenepropanoic acid, 3,4-dimethoxyphenylacetic acid. While the significantly down-regulated 14 metabolites were symmetric dimethylarginine, 1-methyladenosine, citric acid, 5'-deoxyadenosine,

1-(beta-D-ribofuranosyl)-1,4-dihydronicotinamide,

7-methylinosine,

2-keto-6-acetami-

docaproate, 7-aminomethyl-7-carbaguanine, succinyladenosine, benzaldehyde, urothion, 6-hydroxy-5-methoxyindole glucuronide, monobutylphthalate, tetrahydrocortisol. To investigate the degree of change in the tentatively identified biomarkers, the relative intensity of the markers were compared between the APP/PS1 mice and WT mice (Fig. 6). As shown in Fig. 7, clustering heat map analysis of the 24 metabolites revealed the differences of relative value between APP/PS1 mice and WT mice. Metabolic pathways and function analysis Based on the identified endogenous metabolites, a metabolic pathway analysis was performed using MetPA to explore the possible pathways that were affected by APP/PS1 genes. Relevant metabolic pathways which were important for the host response to AD were constructed (Fig.8A.), including pentose and glucuronate interconversions, glyoxylate and dicarboxylate metabolism, starch and sucrose metabolism, citrate cycle, arginine and proline metabolism and tryptophan metabolism. The pentose and glucuronate interconversions, glyoxylate and dicarboxylate metabolism (Impact>0.2) may be acutely perturbed by APP/PS1 genes. Meanwhile, these metabolites 7 ACS Paragon Plus Environment

Journal of Proteome Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 23

with fold change were imported into IPA software to build metabolic networks which is associated with AD (Fig. 8B). The results showed that the level of 5-hydroxyindoleacetic acid was up-regulated as a result of down-regulated N, N-dimethylarginine, which is an inhibitor of nictric oxide synthase. Moreover, it was found that metabolism abnormalities were occurred in the monoaminergic neurotransmitters (Fig. 8C).

DISCUSSION The APP/PS1 mouse model develops some of the biological hallmarks and cognitive decline observed in human Alzheimer with a phenotype characterized by deposition of Aβ plaques starting from the age of four months and deficits in cognitive functions at age of six months.15 In this study, Y maze and Morris water maze were used to test the learning and memory function of mice, and a urinary metabolomics based on UPLC-Q-TOF MS approach was implemented to study metabolites of APP/PS1 transgenic mice at 2 month of age mimicking early stage of AD patients. The results indicated that the urine metabolic profiles were disturbed by APP/PS1 genes prior to any behavioral changes at the preclinical AD stage. 24 differential metabolites were identified and relevant metabolic pathways were disturbed. The possible role of nitric oxide (NO) in the development of neurodegenerative illness is attracting growing attention. 16

L-arginine, asymmetric dimethylarginine (ADMA) and symmetric dimethylarginine (SDMA) are involved in the

NO production pathway. L-arginine is converted to NO by NO synthase (NOS). Unlike ADMA, which directly inhibits NOS, SDMA may inhibit NO synthesis by inhibiting L-arginine uptake in vitro.17 In our work, a reduced level of SDMA in APP/PS1 mice was observed. 4-Guanidinobutanoic acid is formed by the transfer of the guanidine group form arginine to GABA via transamination.18 It has been reported that arginine can be found in humans bio-fluid, including plasma, urine and CSF.19 A large number of arginine accumulations in the body could lead to increased synthesis of 4-Guanidinobutanoic acid, which may impair the nervous system. We also observed 4-Guanidinobutanoic acid was upregulated in the APP/PS1 mice urine samples. The results indicated that the NO production pathway may be disturbed at 2 months of APP/PS1 mice. Studies have also reported that the excessive NO produced by inducible NOS (iNOS) result in an enhancement in the gene expression levels of proteins related to Aβ production, including APP genes and BACE1 genes.20 1-methyladenosine, 5-deoxyadenosine are urinary oxidized nucleosides. They were used as indicators to reveal the association between oxidative stress and neurodegenerative process.21 We observed that the two putatively identified metabolites levels were changed 2-month-old APP/PS1 mice urine samples. 1-(beta-D-ribofuranosyl)-1,4-dihydronicotinamide is the reduced form of nicotinamide riboside. Nicotinamide riboside is a precursor of neuronal nicotinamide adenine dinucleotide (NADH) and a source of vitamin B3.22 Studies showed that nicotinamide riboside may prevent Aβ production in the brain by promoting peroxisome 8 ACS Paragon Plus Environment

Page 9 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

proliferator-activated receptor-γ coactivator-1α-mediated BACE1 ubiquitination and degradation.23 In APP/PS1 mice, there was a decrease in 1-(beta-D-ribofuranosyl)-1,4-dihydronicotinamide levels. Nicotinuric acid is the major catabolic product of nicotinic acid and is also a minor metabolite of fatty acid beta-oxidation. It is involved in nicotinate and nicotinamide metabolism as an endogenous end product, participating in glycolysis, gluconeogenesis and the citric acid cycle.24 It has been reported that the urine nicotinuric acid value of subjects with diabetes was increased and positively correlated with blood pressure, total cholesterol, low-density lipoprotein cholesterol and triacylglycerol.25 As we all know, diabetes increases the incidence of AD. In this study, the level of nicotinuric acid was significantly higher in APP/PS1 mice than WT mice. Tryptophan, phenylalanine and tyrosine, as aromatic amino acids, are the precursors of the central synthesis of neurotransmitters.26 Research has shown that monoaminergic neurotransmitters alterations play a critical role in the course of dementia.27 5-hydroxy-indoleacetic acid (5-HIAA) is a breakdown product of serotonin that is excreted in the urine. A metabolomics study showed that 5-HIAA was increased in both AD and MCI participants.28 It was reported that NO may act as a mediator of 5-HT-evoked secretions.29 In this study, we found that the 5-HIAA level was changed in the urine samples of APP/PS1 mice, which indicated that serotonin metabolic pathway may play a role in the progression of AD. 3,4-Dihydroxyphenylglycol (DHPG) is produced by the metabolism of norepinephrine (NE) by monoamine oxidase (MAO), which was recognized as an estimate of NE clearance. Noradrenergic dysfunction may contribute to cognitive deficits.30 Proteome analysis showed an increase of MAO protein expression in AD.31 Therefore, it is possible that upregulation of MAO in AD could result in the observed increases of DHPG in APP/PS1 mice. Dopamine is also mainly catalyzed by MAO to produce 3,4- dimethoxyphenylacetic acid. Increased DA elimination is associated with cognitive impairment.32 Cellular experiment found that stimulation of dopamine receptors can protect from oligomeric Aβ-induced loss of synaptic plasticity by activation of Src-family tyrosine kinases.33 Collectively, these results suggested that a metabolic abnormality of monoaminergic neurotransmitters may occur as early as two months of age in APP/PS1 mice.

CONCLUSION Alzheimer’s disease, a neurodegenerative disorder, seriously affects human life. It is important to find a reliable and sensitive method to predictive it early. In this study, a metabolomics approach based on ultra-high performance liquid chromatography-mass/mass spectrometry was implemented to study small molecules metabolites present in urine of APP/PS1 mice at age of 2 months. It was found that the urine metabolic profiles were disturbed priors to any behavioral change appears. Multivariate data analysis revealed alterations in 24 metabolites, which could be involved in pentose and glucuronateinterconversions, glyoxylate and dicarboxylate metabolism, starch and sucrose metabolism, citrate cycle, tryptophan metabolism and arginine and proline metabolism, providing a better 9 ACS Paragon Plus Environment

Journal of Proteome Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 23

understanding about the pathogenic mechanism of AD. There is a prominent change in metabolites associated with the NO production pathway and the metabolism of monoaminergic neurotransmitters in APP/PS1 mice at age of 2-month old. This finding demonstrated the potential of a urinary metabolomics approach to discover available biomarkers in APP/PS1 models and may be provide better insights to the pathogenesis of AD.

ASSOCIATED CONTENT Supporting Information Table S1- Differentiating metabolites between APP/PS1 and WT mice, Table S2- Characterization of potential biomarkers in positive and negative mode.

Acknowledgments This work was supported by grants from the Key Program of Natural Science Foundation of State (Grant No. 81430093, 81373930, 81673586, 81302905), National Key Subject of Drug Innovation (Grant No. 2015ZX09101043-005, 2015ZX09101043-011), TCM State Administration Subject of Public Welfare of (Grant No. 2015468004), Natural Science Foundation of Heilongjiang Province of China (H2015038), University Nursing Program for Young Scholars with Creative Talents in Heilongjiang Province (UNPYSCT-2015118).

Competing financial interests The authors declare no competing financial interests.

REFERENCE [1] Owen, J. B.; Di Domenico, F.; Sultana, R.; Perluigi, M., Cini, C.; Pierce, W. M.; Butterfield, D. A. Proteomicsdetermined diffences in the concanavalin-A-fractionated proteome of hippocampus and inferior parietal lobule in subjects with Alzheimer’s disease and mild cognitive impairment: implications for progression of AD. J. Proteome Res. 2009, 8, 471-482. [2] Braak, H.; Braak, E. Neuropathological stageing of Azheimer-related changes. Acta Neuropathol. 1991, 82, 239-259. [3] Arendt, T.; Bigal, V.; Arendt, A.; Tennstedt, A. Loss of neurons in the nucleus basalis of Meynert in Alzheimer’s disease, paralysis agitans and Korsakoff’s Disease. Acta Neuropathol. 1983, 61, 101 -108. [4] Adams, I. M. Structural plasticity of synapses in Alzheimer’s disease. Mol. Neurobiol. 1991, 5, 411-419. [5] Sperling, R. A.; Aisen, P. S.; Beckett, L. A.; Bennett, D. A.; Craft, S.; Fagan, A. M.; Iwatsubo, T., Jack, C. R. Jr.; Kaye, J.; Montine, T. J.; Park, D. C.; Reiman, E. M.; Rowe, C. C.; Siemers, E.; Stern, Y.; Yaffe, K.; Carrillo, M. C.; Thies, B.; Morrison-Bogorad, M.; Wagster, M. V.; Phelps, C. H. Toward defining the preclinical stages of 10 ACS Paragon Plus Environment

Page 11 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

Alzheimer’s disease: recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dementia. 2011, 7, 280-292. [6] Zhao, L.; Chen, T.; Wang, C.; Li, G.; Zhi, W.; Yin, J., Wan, Q.; Chen, L. Atorvastatin improvement of cognitive impairments caused by amyloid β in mice: involvement of inflammatory reaction. BMC Neurol. 2016, 16, 18. [7] Antonios, G.; Borgers, H.; Richard, B. C.; Brauβ, A.; Meiβner, J.; Weggen, S.; Pena, V.; Pillot, T.; Davies, S. L.; Bakrania, P.; Matthews, D.; Brownlees, J.; Bouter, Y.; Bayer, T. A. Alzheimer therapy with an antibody against N-terminal Abeta 4-X and pyroglutamateAbeta 3-X. Sci. Rep. 2015, 5, 177338. [8] Herold, C.; Hooli, B. V.; Mullin, K.; Liu, T.; Roehr, J. T.; Mattheisen, M.; Parrado, A. R.; Bertram, L.; Lange, C.; Tanzi, R. E. Family-based association analyses of imputed genotypes reveal genome-wide significant association of Alzheimer's disease with OSBPL6, PTPRG, and PDCL3. Mol. Psychiatry. 2016, 21, 1608-1612. [9] Melnikova, T.; Savonenko, A.; Wang, Q.; Liang, X.; Hand, T.; Wu, L.; Kaufmann, W. E.; Vehmas, A.; Andreasson K. I. Cycloxygenase-2 activity promotes cognitive deficits but not increased amyloid burden in a model of Alzheimer’s disease in a sex-dimorphic pattern. Neuroscience. 2006, 141, 1149-1162. [10] Nicholson, J. K.; Connelly, J.; Lindon, J. C.; Holmes, E. Metabonomics: a platform for studying drug toxicity and gene function. Nat. Rev. Drug Discovery. 2002, 1, 153-161. [11] González-Domínguez, R.; García-Barrera, T.; Vitorica, J.; Gómez-Ariza, J. L. Application of metabolomics based on direct mass spectrometry analysis for the elucidation of altered metabolic pathways in serum from the APP/PS1 transgenic model of Alzheimer’s disease. J. Pharm. Biomed. Anal. 2015, 107, 378-85. [12] Fukuhara, K.; Ohno, A.; Ota, Y.; Senoo, Y.; Maekawa, K.; Okuda, H.; Kurihara, M.; Okuno, A.; Niida, S.; Saito, Y.; Takikawa, O. NMR-based metabolomics of urine in a mouse model of Alzheimer’s disease: identification of oxidative stress biomarkers. J. Clin. Biochem. Nutr. 2013, 52, 133-138. [13] Zhao, Q.; Niu, Y.; Matsumoto, K.; Tsuneyama, K; Tanaka, K.; Miyata, T.; Yokozawa, T. Chotosan ameliorates cognitive and emotional deficits in an animal model of type 2 diabetes: possible involvement of cholinergic and VEGF/PDGF mechanisms in the brain. BMC Complementary Altern. Med. 2012,12:188. [14] Guo, C.; Wang, T.; Zheng, W.; Shan, Z. Y.; Teng, W. P.; Wang, Z. Y. Intranasal deferoxamine reverses ironinduced memory deficits and inhibits amyloidogenic APP processing in a transgenic mouse model of Alzheimer’s disease. Neurobiol. Aging. 2013, 34:562- 575. [15] Lee, S. H.; Kim, I.; Chung, B. C. Increased urinary level of oxidized nucleosides in patients with mild-to moderate Alzheimer’s disease. Clin. Biochem. 2007, 40, 936-938. [16] Hannibal, L. Nitric oxide homeostasis in neurodegenerative disease. Curr. Alzheimer Res. 2016, 13, 135-49. [17] Mommersteeg, P. M.; Schoemaker, R. G.; Eisel, U. L.; Garrelds, I. M.; Schalkwijk, C. G.; Kop, W. J. Nitric oxide dysregulation in patients with heart failure: the association of depressive symptoms with L-arginine, asymmetric 11 ACS Paragon Plus Environment

Journal of Proteome Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 23

dimethylarginine, symmetric dimethylarginine, and isoprostane. Psychosom. Med. 2015, 77, 292-302. [18] Jasen, E. E.; Verhoeven, N. M.; Jakobs, C.; Schulze, A.; Senephansiri, H.; Gupta, M.; Snead, O. C.; Gibson, K. M. Increased guanidino species in murine and human succinate semialdehyde dehydrogenase (SSADH) deficiency. Biochim. Biophys. Acta. 2006, 1762, 494-498. [19] Wang, S. Y.; Wang, Y.; Jin, X. W.; Zhang, Y.; Chen, J. S.; Ma, W. W.; Wu, Y. H. Wang, D. C. A urinary metabolomics study of rats after the exposure to acrylamide by ultra performance liquid chromatography coupled with quadrupole time-of-flight tandem mass spectrometry. Mol. Biosyst. 2015, 11, 1146-1155. [20] Nagai, N.; Ito, Y.; Shibata, T.; Kubo, E.; Sasaki, H. A positive feedback loop between nitric oxide and amyloid β (1-42) accelerates mitochondrial damage in human lens epithelial cells. Toxicology. 2017, 381, 19-30. [21] Lee, S. H.; Kim, I.; Chung, B. C. Increased urinary level of oxidized nucleosides in patients with mild-to –moderate Alzheimer’s disease. Clin. Biochem. 2007, 40, 936-938. [22] Chi, Y.; Sauve, A. A. Nicotinamideriboside, a trace nutrient in foods, is a vitamin B3 with effects on energy metabolism and neuroprotection. Curr. Opin. Clin. Nutr. Metab. Care. 2013, 16, 657-661. [23] Gong, B.; Pan, Y.; Vempati, P.; Zhao, W.; Knable, L.; Ho, L; Wang, J.; Sastre, M.; Ono, K.; Sauve, A. A.; Pasinetti, G. M. Nicotinamideriboside restores cognition through an upregulation of proliferator-activated receptor-γ coactivator 1α regulated β-secretase 1 degradation and mitochondrial gene expression in Alzheimer's mouse models. Neurobiol. Aging. 2013, 34, 1581-1588. [24] Shen, C.; Sun, Z.; Chen, D.; Su, X.; Jiang, J.; Li, G.; Lin, B.; Yan, J. Developing urinary metabolomic signatures as early bladder cancer diagnostic markers. OMICS. 2015, 19, 1-11. [25] Huang, C. F.; Cheng, M. L.; Fan, C. M.; Hong, C. Y. Shiao, M. S. Nicotinuric acid: a potential marker of metabolic syndrome through a metabolomics-based approach. Diabetes Care. 2013, 36, 1729-1731. [26] Birkl, P.; Franke, L.; Bas Rodenburg, T.; Ellen, E.; Harlander-Matauschek, A. A role for plasma aromatic amino acids in injurious pecking behavior in laying hens.Physiol. Behav. 2017, 175, 88-96. [27] Vermeiren, Y.; Le Bastard, N.; Van Hemelrijck, A.; Drinkenburg, W. H.; Engelborghs, S.; De Deyn, P. P. Behavioral correlates of cerebrospinal fluid amino acid and biogenic amine neurotransmitter alterations in dementia. Alzheimer’s Dementia. 2013, 9, 488-498. [28] kaddurah-Daouk, R.; Zhu, H.; Sharma, S.; Bogdanov, M.; Rozen, S. G.; Matson, W.; Oki, N. O.; Motsinger-Reif, A. A.; Churchill, E.; Lei, Z.; Appleby, D.; King, M. A.; Trojianowski, J. Q.; Doraiswamy, P. M.; Arnold, S. E. Pharmacometabolomics Res, N. Alterations in metabolic pathways and networks in Alzheimer’s disease. Transl. Psychiatry. 2013, DOI: 10.1038/tp.2013.18. [29] Yazar, A.; Büyükafpar, K.; Polat, G.; Pata, C.; Kanýk, A.; Tiftik, E. N.; Bağdatoğlu, O. The urinary 5-hydroxyindole acetic acid and plasma nitric oxide levels in irritable bowel syndrome: a preliminary study. Scott. 12 ACS Paragon Plus Environment

Page 13 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

Med. J. 2005, 50, 27-29. [30] Friedman, J. I.; Adler, D. N.; Davis, K. L. The role of norepinephrine in the pathophysiology of congnitvie disorders: potential applications to the treatment of cognitive dysfunction in schizophrenia and Alzheimer’s disease. Biol. Psychiatry. 1999, 46, 1243-1252. [31] Zellner, M.; Baureder, M.; Rappold, E.; Bugert, P.; Kotzailias, N.; Babeluk, R.; Baumgartner, R.; Attems, J.; Gerner, C.; Jellinger, K. Roth, E.; Oehler, R.; Umlauf, E. Comparative platelet proteome analysis reveals an increase of monoamine oxidase-B protein expression in Alzheimer’s disease but not in non-demented Parkinson’s disease patients. J. Proteomics. 2012, 75, 2080-2092. [32] Moreno-Castilla, P.; Rodriguez-Duran, L. F.; Guzman-Ramos, K.; Barcenas-Femat, A.; Escobar, M. L.; Bermudez-Rattoni, F. Dopaminergic neurotransmission dysfunction induced by amyloid-β transforms cortical long-term potentiation into long-term depression and produces memory impairment. Neurobiol. Aging. 2016, 41, 187 -199. [33]Xiang, P. Y.; Janc, O.; Grochowska, K. M.; Kreutz, M. R. Reymann, K. G. Dopamine agonists rescue Aβ-induced LTP impairment by Scr-family tyrosine kinases. Neurobiol. Aging. 2016, 40, 98-102.

13 ACS Paragon Plus Environment

Journal of Proteome Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 23

Figure legends Fig.1. Y maze and Morris water maze performance of 2-month old APP/PS1 miceandwild-type mice. (A) Schematic drawings of the Y-maze. The maze was surrounded by different spatial cues. (B) %Time spent in a novel arm were analyzed. Each data column indicates the mean±S.D. Learning performance of mice was analyzed by escape latency (C). (D) %Time spent in the target quadrant were analyzed. Each datum presents the mean±S.D. Fig.2. Score plots of PCA model for metabolomic data from urine samples. Wild-type mice (blue diamonds) , APP/PS1 mice (red triangle) in positive mode (A) and negative mode (B). Fig.3. Metabolic profiling analysis of the UPLC/MS spectra of urine samples. 3D-PCA score plots in positive ion mode (A) and negative ion mode (B). Black dot represents wild-type mice and red dot represents APP/PS1 mice. Urinary BPI chromatograms of APP/PS1 miceandwild-type micein positive ion mode (C) andnegative ion mode (D). Fig.4. OPLS-DA analy for metabolomic data from urine samples. Score plot of OPLS-DA in positive mode (A) and in negative mode (B). VIP-score plot of OPLS-DA in positive mode (C) and negative mode (D). Blue diamond represents wild-type mice, and red triangle represents APP/PS1 mice. Fig.5. High resolution M/MS spectra and fragmentation pathways of 6-Hydroxy-5-methoxyindole glucuronide in positive mode. The mass spectrum at MS/MS mode (A), proposed fragmentation pathways of 6-hydroxy-5-methoxyindoleglucuronide (B). Fig.6.Comparison of the relative intensity of the potential biomarkers in APP/PS1 transgenic mice and wide-type mice. Fig.7. Heat map analysis of 24 differentiating metabolites between APP/PS1 mice and wide type mice. The degree of change marked with different colors-red represents up-regulation and green indicates down-regulation. Each row represents an individual sample, and each column represents a metabolite. Fig.8. Network analysis of biomarkers. (A) Metabolism pathway in APP/PS1 mice. The map was generated using MetPA (http://metpa.metabolomics.ca./MetPA/faces/Home.jsp). (a), glyoxylate and dicarboxylate metabolism; (b), pentose and glucronateinterconversions; (c), starch and sucrose metabolism; (d), citrate cycle; (e), arginine and proline metabolism; (f), tryptophan metabolism. (B) Biomarkers network analysis based on IPA. (C) Biosynthetic and metabolic pathway of biogenic amine neurotransmitter. The red represents an increase in content, while the green represents a decrease in content.

14 ACS Paragon Plus Environment

Page 15 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

Fig.1

15 ACS Paragon Plus Environment

Journal of Proteome Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 23

Fig.2

16 ACS Paragon Plus Environment

Page 17 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

Fig.3

17 ACS Paragon Plus Environment

Journal of Proteome Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 23

Fig.4

18 ACS Paragon Plus Environment

Page 19 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

Fig.5

19 ACS Paragon Plus Environment

Journal of Proteome Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 23

Fig.6

20 ACS Paragon Plus Environment

Page 21 of 23

Journal of Proteome Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Fig.7

21 ACS Paragon Plus Environment

Journal of Proteome Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 23

Fig.8

22 ACS Paragon Plus Environment

Page 23 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

For TOC only

23 ACS Paragon Plus Environment