Metabolomic screening of regional brain alterations in the APP/PS1 transgenic model of Alzheimer's disease by direct infusion mass spectrometry

Elsevier Editorial System(tm) for Journal of Pharmaceutical and Biomedical Analysis Manuscript Draft Manuscript Number: JPBA-D-14-00875R1 Title: METABOLOMIC SCREENING OF REGIONAL BRAIN ALTERATIONS IN THE APP/PS1 TRANSGENIC MOUSE MODEL OF ALZHEIMER'S DISEASE BY DIRECT INFUSION MASS SPECTROMETRY Article Type: Full Length Article Keywords: Metabolomics; APP/PS1 mice; brain regions; direct infusion mass spectrometry; Alzheimer's disease Corresponding Author: Prof. JOSE LUIS GOMEZ-ARIZA, Ph.D. Corresponding Author's Institution: UNIVERSITY OF HUELVA First Author: Raul Gonzalez-Dominquez, Ms Order of Authors: Raul Gonzalez-Dominquez, Ms; Tamara Garcia-Barrera, Ph. Dr.; Javier Vitorica, Ph. Dr.; JOSE LUIS GOMEZ-ARIZA, Ph.D. Abstract: The identification of pathological mechanisms underlying to Alzheimer's disease is of great importance for the discovery of potential markers for diagnosis and disease monitoring. In this study, we investigated regional metabolic alterations in brain from the APP/PS1 mice, a transgenic model that reproduces well some of the neuropathological and cognitive deficits observed in human Alzheimer's disease. For this purpose, hippocampus, cortex, cerebellum and olfactory bulbs were analyzed using a high-throughput metabolomic approach based on direct infusion mass spectrometry. Metabolic fingerprints showed significant differences between transgenic and wild-type mice in all brain tissues, being hippocampus and cortex the most affected regions. Alterations in numerous metabolites were detected including phospholipids, fatty acids, purine and pyrimidine metabolites, acylcarnitines, sterols and amino acids, among others. Furthermore, metabolic pathway analysis revealed important alterations in homeostasis of lipids, energy management, and metabolism of amino acids and nucleotides. Therefore, these findings demonstrate the potential of metabolomic screening and the use of transgenic models for understanding pathogenesis of Alzheimer's disease. Suggested Reviewers: Dear Editor, Please find enclosed the manuscript “METABOLOMIC SCREENING OF REGIONAL BRAIN ALTERATIONS IN THE APP/PS1 TRANSGENIC MOUSE MODEL OF ALZHEIMER’S DISEASE BY DIRECT INFUSION MASS SPECTROMETRY" by R. González-Domínguez, T. García-Barrera, J. Vitorica and J.L. Gómez-Ariza, to be published in Journal of Proteomics. This work considers the application for the first time of direct infusion electrospray mass spectrometry for the study of metabolic abnormalities in different brain regions of the APP/PS1 transgenic mice, including cortex, hippocampus, cerebellum and olfactory bulbs. Multivariate statistics demonstrated the potential of high-throughput fingerprinting for the discrimination between transgenic animals and wild-type controls, and indicated that pathological processes are widespread and not only affect to hippocampus and cortex, primary targets in Alzheimer’s disease. Furthermore, numerous metabolites were identified as potential markers of AD-type disorders in these mice, which may contribute to deepen into underlying pathological mechanisms related to neurodegenerative processes. Therefore, these results highlight the importance of transgenic models for the study of underlying pathological mechanisms in AD brain, a sample not readily accessible in human investigations. Sincerely yours, José Luis Gómez Ariza Cover Letter Combined Reviewers' comments: Compound Identification: 1. Direct infusion mass spectrometry lacks of the information provided by retention time and the resolution of different signals for different compounds. In this scenario the authors say: (line 138) "For this, the peak search was done with a mass tolerance of 0.1Da, and a minimum response of 10 counts was considered for filtering". The above error in mass means working with nominal mass more than exact mass and the level of signal means accepting even noise as a signal. Any information obtained under those conditions is doubtful. It is true that they can confirm some lipids through positive and negative ionization modes and with the fragments, but any other compound is in my opinion just an educated guess. The “mass tolerance of 0.1Da” mentioned in line 138 refers to the m/z width used for data filtering (not mass accuracy). Data filtering in metabolomics based on direct infusion with TOF-MS is usually performed following the procedure employed in our manuscript. Hansen et al. described that the optimal m/z width for data binning during this filtering step must be “determined by the mass resolution in such a way that two close, but separate mass peaks will not be mixed together”, by applying a weighted polynomial filtering method (Metabolomics 2007 3:41). If this m/z value is increased “the smoothing will fit to the underlying spectrum resulting in over-fitting”, while decreasing this width the “smaller peaks are gradually removed”. Thereby, mass spectra from DIMS are usually binned in intervals of 0.01-0.5 Da (see Analyst 2010 135:2970; Int J Mass Spectrometry 2012 309:200; Metabolomics 2010 6:156; Int J Genomics 2014 894296). For this reason, we selected m/z width = 0.1 Da. Furthermore, peak selection must be accompanied by a “noise elimination” step. In this work the noise level was set to 10 counts because several compounds presented very small signal intensities (principally at low m/z values, determined empirically from experimental spectra), although the use of this nonrestrictive cutoff obviously leads to the inclusion of numerous spurious signals in the final data matrix. However, given that spurious signals are randomly distributed across all the samples, these peaks do not influence significantly subsequent statistical analyses. Taking all this into account, the abovementioned sentence (line 138) has been re-written to clarify the data filtering step. Latter on the authors use all types of mass errors: 2. Line 156: "Potential biomarkers were identified matching the experimental accurate mass". Which experimental accurate mass with 0.1 Da error? And then: Line 160 …"characteristic ions in positive ionization mode at m/z 184.07, 104.10 and 86.09, and two typical fragments…"; they used two decimal figures, while in Tables they used three. As stated in the response to the first question, 0.1Da is the m/z width employed for data binning (not mass error). All m/z values from the entire manuscript have been revised to display three decimals. Chemometric models: 3. Line 147: Quality of the models was assessed by the R2 and Q2 values, supplied by the software, which provided information about the class separation and predictive power of the model, respectively. Looking at the poor clustering of groups in PCA and being an animal model, the separation of groups was very weak, which means that PLS-DA could be overfitted; the chemometric model should be validated in any of the different ways: permutation test, leaving 1/3 out, validating OPLS-DA…. Previous metabolomic works in different transgenic models of AD also reported that PCA is not able to separate study groups, being necessary the application of supervised methods such as PLS-DA (J Clin Biochem Nutr 2013 52:133, J Proteome Res 2008 7:3678). Anyway, PLS-DA models were validated using permutation tests. Permuted models showed lower R2 and Q2 values than original models, indicating that models are not overfitted. This has been mentioned in the revised manuscript Statistics: 4. Line 151: potential biomarkers were selected according to the Variable Importance in the Projection, or VIP (a weighted sum of squares of the PLS weight, which indicates the importance of Response to Reviewers the variable in the model), considering only variables with VIP values higher than 1.5, indicative of significant differences among groups. VIP values should be given in the Tables, in addition Jack Knife intervals should be calculated to prove the statistical significance of variables. Have the authors calculated p values? Potential biomarkers were selected according to the VIP-plots with confidence intervals derived from jack-knifing, and then were statistically validated using t-test (this has been corrected in the manuscript). VIP and p-values have been appended to tables 2-4. Other comments 5. Are there differences between male and female according to the disease? When p-values of potential markers listed in Tables 2-4 are compared with p-values according to gender for these metabolites, we can observe that the effect of gender is much less important that the disease state (see attached figure, for data from hippocampus). Only a few metabolites presented a statistically different trend between male and female mice (e.g. inosine, valine), in line with findings reported by van Duijn et al. (J Alzheimers Dis 2013 34:1051). However, the evaluation of gender-specific metabolic changes in the APP/PS1 model was out of the scope of this manuscript. 6. Considering that this is an analytical journal the tables should include the ions (adducts...) that lead to the identification. Ions detected in MS/MS experiments have been appended to Tables 2-4. Some minor mistakes in the language 7. Line 298: "transgenic mice exhibitS altered…" It was corrected. -7 -6 -5 -4 -3 -2 -1 0 p-value (log10) metabolites gender disease *Graphical Abstract METABOLOMIC SCREENING OF REGIONAL BRAIN ALTERATIONS IN THE APP/PS1 TRANSGENIC MODEL OF ALZHEIMER’S DISEASE BY DIRECT INFUSION MASS SPECTROMETRY Raúl González-Domínguez, Tamara García-Barrera, Javier Vitorica, José Luis Gómez-Ariza  Direct infusion mass spectrometry allows a comprehensive brain metabolomic analysis  The APP/PS1 mice exhibited an abnormal neurochemical profile compared to controls  These failures affected hippocampus, cortex, cerebellum and olfactory bulbs  Pathway analysis revealed multiple significant impairments in brain metabolism *Highlights (for review) 1 METABOLOMIC SCREENING OF REGIONAL BRAIN ALTERATIONS IN THE APP/PS1 1 TRANSGENIC MODEL OF ALZHEIMER’S DISEASE BY DIRECT INFUSION MASS 2 SPECTROMETRY 3 4 a,b,cRaúl González-Domínguez, a,b,cTamara García-Barrera*, d,e,fJavier Vitorica, a,b,cJosé Luis Gómez-5 Ariza* 6 7 aDepartment of Chemistry and CC.MM. Faculty of Experimental Sciences. University of Huelva. Campus 8 de El Carmen. 21007 Huelva. SPAIN; bCampus of Excellence International ceiA3. University of Huelva. 9 SPAIN; cResearch Center of Health and Environment (CYSMA). University of Huelva. Campus de El 10 Carmen. 21007 Huelva. SPAIN; dDepartment Bioquímica, Bromatologia, Toxicología y Medicina Legal, 11 Faculty of Pharmacy, University of Seville. 41012 Seville. SPAIN, eCentro de Investigación Biomédica 12 en Red sobre Enfermedades Neurodegenerativas (CIBERNED). 41013 Seville. SPAIN, fInstituto de 13 Biomedicina de Sevilla (IBiS)–Hospital Universitario Virgen del Rocío/CSIC/University of Seville. 14 41013 Seville. SPAIN 15 16 17 Corresponding authors: 18 Prof. J.L Gómez Ariza, Tel.: +34 959 219968, fax: +34 959 219942, e-mail: [email protected] 19 Dr. T. García-Barrera, Tel.: +34 959 219962, fax: +34 959 219942, e-mail: ta[email protected] 20 21 22 23 ABSTRACT 24 The identification of pathological mechanisms underlying to Alzheimer’s disease is of great importance 25 for the discovery of potential markers for diagnosis and disease monitoring. In this study, we investigated 26 regional metabolic alterations in brain from the APP/PS1 mice, a transgenic model that reproduces well 27 some of the neuropathological and cognitive deficits observed in human Alzheimer’s disease. For this 28 purpose, hippocampus, cortex, cerebellum and olfactory bulbs were analyzed using a high-throughput 29 metabolomic approach based on direct infusion mass spectrometry. Metabolic fingerprints showed 30 *Revised Manuscript Click here to view linked References 2 significant differences between transgenic and wild-type mice in all brain tissues, being hippocampus and 31 cortex the most affected regions. Alterations in numerous metabolites were detected including 32 phospholipids, fatty acids, purine and pyrimidine metabolites, acylcarnitines, sterols and amino acids, 33 among others. Furthermore, metabolic pathway analysis revealed important alterations in homeostasis of 34 lipids, energy management, and metabolism of amino acids and nucleotides. Therefore, these findings 35 demonstrate the potential of metabolomic screening and the use of transgenic models for understanding 36 pathogenesis of Alzheimer’s disease. 37 38 39 40 41 42 43 44 45 KEYWORDS 46 Metabolomics, APP/PS1 mice, brain regions, direct infusion mass spectrometry, Alzheimer’s disease 47 48 49 50 51 52 53 54 55 56 57 58 59 60 3 61 62 1. INTRODUCTION 63 Numerous transgenic mouse models of Alzheimer’s disease (AD) have been developed for understanding 64 disease pathology and testing potential therapies. The ideal model should show the full range of clinical 65 and pathological features associated with AD, including cognitive and behavioral deficits, amyloid 66 plaques, neurofibrillary tangles and neurodegeneration [1]. However, no existing model exhibits all these 67 features, but each one has unique pathologies that provide insights into disease mechanisms. One of the 68 models most extensively used is the double transgenic line APP/PS1, expressing the Swedish mutation of 69 -amyloid precursor protein together with deleted presenilin 1 in exon 9, which reproduces well some of 70 the neuropathological and cognitive deficits observed in AD, with a phenotype characterized by early 71 amyloid deposits and behavioral impairment [2]. Behavioral assessments have been traditionally used to 72 confirm cognitive deficits in these transgenic animals [3], but they are tedious and may suffer from high 73 individual variability. Thus, the pathophysiological status of these models is better evaluated by analysis 74 of brain tissue samples to detect possible biomarkers. In this context, metabolomics may present a high 75 potential for identifying neurochemical changes involved in pathological mechanisms occurring in AD, 76 since provides a comprehensive overview of the status of organism reflecting the interactions between 77 genes, proteins and the environment [4]. Non-invasive metabolic profiling by means of in vivo magnetic 78 resonance spectroscopy (MRS) has been widely applied to AD studies in different transgenic models [5-79 7]. The most consistent findings observed using this approach are the decrease of N-acetyl aspartate 80 (NAA) levels, a biomarker for neuronal integrity, and the increase of myo-inositol (mIns), which is 81 thought to be a marker for osmotic stress or astrogliosis. Additionally, changes in other metabolites such 82 as glutamate, creatine and choline-containing compounds were also found. On the other hand, in vitro 83 nuclear magnetic resonance (NMR)-based metabolomic investigations of postmortem brain have been 84 also proposed, providing a wider range of measurable metabolites compared to in vivo MRS [8-10]. 85 However, the characterization of regional metabolomic perturbations instead of overall brain changes 86 may be of greater interest in order to investigate the impact of disease on different brain regions and 87 determine the most affected ones in AD-mice. In this sense, only a few authors have previously 88 performed a comparative metabolomic study in different brain regions using NMR, demonstrating that 89 the hippocampus and temporal cortex are the most sensitive regions to disease, but other tissues are also 90 10 type of fatty acid linked to the molecular moiety, as recently described for human AD where membrane 269 destabilization processes were associated with imbalances in the levels of saturated/unsaturated fatty 270 acids contained in the structure of phospholipids [33]. Thereby, major changes observed in this study 271 corresponded to reduced levels of PUFA-containing phospholipids, principally phospatidylcholines (PC) 272 and inositols (PI) containing docosahexaenoic acid (DHA), and most phosphatidylethanolamines (PE) 273 and plasmenylethanolamines (PPE), which is in accordance with previous studies in brain from transgenic 274 mice of AD [38-40]. However, a parallel accumulation of specific compounds was also observed in 275 phospholipids derived from docosapentaenoic and docosatetraenoic acids, as well as different stearoyl-276 arachidonoyl-phospholipids, not previously described to our knowledge. In first place, elevated content of 277 docosapentaenoic and docosatetraenoic acids in the brain phospholipid pool may be correlated to 278 peroxisomal dysfunction, given that they are intermediates for the biosynthesis of long chain 279 polyunsaturated fatty acids such as DHA in peroxisomes. In this sense, substantial peroxisome-related 280 alterations have been previously described in AD brains, inducing the accumulation of very long chain 281 fatty acids and deficits of plasmalogens and DHA [41]. Moreover, the increase of stearoyl-arachidonoyl-282 phospholipids (PC, PE, PI) denotes a profound membrane remodeling in the APP/PS1 mice as it is one of 283 the most abundant species on the brain [42], being cortex the most affected region. Therefore, alterations 284 in phospholipids metabolism appears to have a multifactorial origin involving over-activation of 285 phospholipases, peroxisomal dysfunction and abnormal fatty acid composition of phospholipids. 286 287 Alternatively, deficits for most acylcarnitines in brain of the APP/PS1 mice (Table 3) might indicate a 288 perturbed lipid metabolism, also contributing to the accumulation of free fatty acids. Lower levels of L-289 carnitine in brain (Table 3) have been previously reported in AD patients [43], together with altered 290 expression of several related enzymes such as decreased carnitine acetyltransferase activity [44], or over-291 expressed hydroxyacyl-coenzyme A dehydrogenase (HADHA) [45] and short chain 3-hydroxyacyl-CoA 292 dehydrogenase (SCHAD) [46]. These results suggest a perturbed transport of fatty acids into the 293 mitochondria for -oxidation, which together with the increased rate of phospholipids degradation finally 294 induces the accumulation of free fatty acids. Conversely, levels of propionyl-carnitine and propionic acid 295 were increased in brain samples, indicating a specific disturbance in propionate metabolism. In this sense, 296 Cuadrado-Tejedor et al. recently found a differential expression of propionyl-CoA carboxylase in 297 hippocampus of the Tg2576 mouse [47], which could cause the propionic acidemia observed in this 298 11 study. This abnormal catabolism of lipids was accompanied by other impairments in energy metabolism, 299 regarding increased brain levels of pyruvate and alanine (Table 4). Elevation of pyruvate level in AD 300 brain has been previously associated with a decreased rate of carbohydrate utilization by the neuron in 301 response to drastic reductions in the activities of enzymes involved in pyruvate oxidation [48]. In the 302 same way, increased concentration of alanine, produced in the body from the conversion of pyruvate, may 303 also indicate a change in the carbohydrate metabolism of the brain [10-11,17]. Taking all this into 304 account, we can conclude that several cellular impairments occur in brain of the APP/PS1 mice in direct 305 relation to the accumulation of free fatty acids, involving multiple alterations in phospholipids 306 metabolism and different energetic pathways, as schematized in Fig. 3A. 307 308 There is also growing evidence for the involvement of nucleotide metabolism in different 309 neurodegenerative mechanisms in Alzheimer’s disease, in accordance with alterations found for 310 numerous purine and pyrimidine metabolites in all brain regions of the APP/PS1 mice (Table 4). Elevated 311 adenosine monophosphate deaminase activity has been identified in AD brains, provoking accelerated 312 degradation of AMP and over-production of ammonia leading to hyperammonemia [49]. In addition, the 313 decrease of AMP levels may have important consequences in cellular energy homeostasis, given that it 314 plays a central role in glucose and lipid metabolism through the AMP-activated protein kinase (AMPK), 315 which is known to be decreased in AD brain [50]. Moreover, purinergic signaling also appears to play a 316 role in the development of AD, with an up-regulation of adenosine receptors in the frontal cortex of 317 affected brains [51], as well as a redistribution of these receptors, with a higher activity in neurons 318 affected by β amyloid deposition or hyperphosphorilation of τ protein [52]. Dysregulation of pyrimidine 319 metabolism, with decreased uridine monophosphate and increased uracil (Table 4), could reflect reduced 320 synaptic plasticity and neuronal deficits due to decreased synthesis of phosphatidylcholines via the 321 Kennedy cycle [53]. Finally, several studies have also implicated oxidative stress in abnormal metabolism 322 of purines and pyrimidines in AD, demonstrating an increase of oxidized DNA bases in brain [54-55], 323 such as 8-hydroxyadenine, 8-hydroxyguanine or FAPy-adenine, as in our metabolomic screening test. 324 Therefore, metabolism of purines and pyrimidines highlights as a candidate pathway for the search of 325 potential markers of pathological processes occurring in the APP/PS1 transgenic mice (Fig. 3B). 326 327 12 Decreased levels of cholesterol, cholesterol sulfate, deoxycholic (DCA) and taurodeoxycholic (TCA) 328 acids were also found in hippocampus and cortex (Table 4). It has been previously demonstrated that the 329 APP/PS1 transgenic mice exhibit altered cholesterol metabolism, resulting in reduced content of 330 cholesterol in brain [38,56] and generating serious alterations of the physicochemical structure of lipid 331 rafts. The same trend was observed for cholesterol sulfate, a component of cell membranes where it plays 332 a stabilizing role [57], not previously described to our knowledge in AD research. Finally, reductions of 333 brain deoxycholic and taurocholic acids (or isomers) could be behind different neuropathological 334 conditions, given that ursodeoxycholic acid (UDCA) and tauroursodeoxycholic acid (TUDCA) have been 335 demonstrated to be potent inhibitors of apoptosis [58], and they present neuroprotective action against 336 amyloid deposition [59]. 337 338 Other potential markers found by metabolomic fingerprinting of brain tissues could be related to failures 339 in neurotransmitter systems, including deficits in tyrosine, dopamine and aspartate, as well as increased 340 N1-acetylspermidine (Table 4). Dopamine is a neurotransmitter derived from the amino acid tyrosine, 341 commonly linked to Parkinson’s disease. However, disturbances in the biosynthesis of monoaminergic 342 neurotransmitters and their precursors have been also reported in AD subjects [60-61]. On the other hand, 343 aspartate is an excitatory neurotransmitter that, similarly to glutamate, usually presents lower 344 concentrations in AD brain [18] and cerebrospinal fluid [62]. Finally, amyloid beta deposition is known to 345 up-regulate polyamine metabolism in Alzheimer’s disease by increasing ornithine decarboxylase activity 346 and polyamine uptake [63], leading to altered levels of polyamines in brain. In this context, Inoue et al. 347 found an abnormal increase of N1-acetylspermidine and other polyamines in AD brains, which was 348 associated with N-methyl-D-aspartate (NMDA) receptor excitotoxicity [64], confirming our metabolomic 349 findings. The significant reduction in urea levels (Table 4) pointed to a perturbation of the urea cycle, 350 responsible for controlling ammonia concentrations in the organism. In Alzheimer’s disease, the 351 alteration of this pathway has been previously demonstrated on the basis of altered levels of expression in 352 different enzymes and the corresponding genes [65]. To conclude changes observed in homocarnosine 353 and glutathione, important antioxidants involved in the defense of the central nervous system, could be 354 related to oxidative damage in brain. Reduced levels of homocarnosine (Table 4) have been previously 355 described in brains of patients with Alzheimer's disease [66], as well as other related dipeptides such as 356 carnosine [61]. On the other hand, levels of glutathione were surprisingly increased in different brain 357 13 regions of the APP/PS1 mice (Table 4), although most previous studies reported reduced content of this 358 antioxidant during AD development in both humans and transgenic models [8,67]. However, Adams et al. 359 also described an analogous increase in AD brain, proposing that a compensatory elevation of glutathione 360 must occur against cellular damage produced by oxidative stress [68]. 361 362 5. CONCLUSIONS 363 The potential of direct infusion mass spectrometry for metabolomic analysis of brain samples has been 364 demonstrated in this study, in terms of non-targeted metabolite coverage, rapidity of analysis and, 365 consequently, high-throughput screening capability. This approach was employed for the investigation of 366 regional metabolic abnormalities in brain of transgenic APP/PS1 mice of Alzheimer’s disease compared 367 with wild-type control mice. Major differences were observed in hippocampus and cortex, primary brain 368 targets in Alzheimer’s disease, but cerebellum and olfactory bulbs were also affected to a lesser extent. 369 Furthermore, these metabolic alterations could be linked to different pathways associated with 370 pathological mechanisms occurring in the APP/PS1 mice, such as impaired metabolism of fatty acids and 371 phospholipids, bioenergetic failures, altered metabolism of purines and pyrimidines, changes in 372 neurotransmission or oxidative stress. Therefore, these results highlight the importance of transgenic 373 models for the study of underlying pathological mechanisms in AD brain, a sample not readily accessible 374 in human investigations. 375 376 Acknowledgements. This work was supported by the projects CTM2012-38720-C03-01 from the 377 Ministerio de Ciencia e Innovación and P008-FQM-3554 and P009-FQM-4659 from the Consejería de 378 Innovación, Ciencia y Empresa (Junta de Andalucía). Raúl González Domínguez thanks the Ministerio de 379 Educación for a predoctoral scholarship. 380 381 REFERENCES 382 [1] Hall AM, Roberson ED. Mouse models of Alzheimer’s disease. Brain Res Bull 2012;88:3-12. 383 [2] Malm T, Koistinaho J, Kanninen K. Utilization of APPswe/PS1dE9 transgenic mice in research of 384 Alzheimer's disease: Focus on gene therapy and cell-based therapy applications. Int J Alzheimers Dis 385 2011;2011:517160. 386 14 [3] Filalia M, Lalonde R. Age-related cognitive decline and nesting behavior in an APPswe/PS1 bigenic 387 model of Alzheimer’s disease. Brain Res 2009;1292:93-9. 388 [4] Lindon JC, Holmes E, Nicholson JK. Metabonomics and its role in drug development and disease 389 diagnosis. Expert Rev Mol Diagn 2004;4:189-99. 390 [5] Oberg J, Spenger C, Wang FH, Andersson A, Westman E, Skoglund P, Sunnemark D, Norinder U, 391 Klason T, Wahlund LO, Lindberg M. Age related changes in brain metabolites observed by 1H MRS in 392 APP/PS1 mice. Neurobiol Aging 2008;29:1423-33. 393 [6] Xu W, Zhan Y, Huang W, Wang X, Zhang S, Lei H. Reduction of hippocampal N-acetyl aspartate 394 level in aged APPSwe/PS1dE9 transgenic mice is associated with degeneration of CA3 pyramidal 395 neurons. J Neurosci Res 2010;88:3155-60. 396 [7] Chen SQ, Cai Q, Shen YY, Wang PJ, Teng GJ, Zhang W, Zang FC. Age-related changes in brain 397 metabolites and cognitive function in APP/PS1 transgenic mice. Behav Brain Res 2012;235:1-6. 398 [8] Dedeoglu A, Choi JK, Cormier K, Kowall NW, Jenkins BG. Magnetic resonance spectroscopic 399 analysis of Alzheimer’s disease mouse brain that express mutant human APP shows altered 400 neurochemical profile. Brain Res 2004;1012:60-5. 401 [9] Forster DM, James MF, Williams SR. Effects of Alzheimer’s disease transgenes on neurochemical 402 expression in the mouse brain determined by 1H MRS in vitro. NMR Biomed 2012;25:52-8. 403 [10] Graham SF, Holscher C, McClean P, Elliott CT, Green BD. 1H NMR metabolomics investigation of 404 an Alzheimer’s disease (AD) mouse model pinpoints important biochemical disturbances in brain and 405 plasma. Metabolomics 2013;9:974-83. 406 [11] Salek RM, Xia J, Innes A, Sweatman BC, Adalbert R, Randle S, McGowan E, Emson PC, Griffin 407 JL. A metabolomic study of the CRND8 transgenic mouse model of Alzheimer’s disease. Neurochem Int 408 2010;56:937-43. 409 [12] Woo DC, Lee SH, Lee DW, Kim SY, Kim GY, Rhim HS, Choi CB, Kim HY, Lee CU, Choe BY. 410 Regional metabolic alteration of Alzheimer’s disease in mouse brain expressing mutant human APP-PS1 411 by 1H HR-MAS. Behav Brain Res 2010;211:125-31. 412 [13] Lalande J, Halley H, Balayssac S, Gilard V, Déjean S, Martino R, Francés B, Lassalle JM, Malet-413 Martino M. 1H NMR metabolomic signatures in five brain regions of the AβPPswe Tg2576 mouse model 414 of Alzheimer's disease at four ages. J Alzheimers Dis 2014;39:121-43. 415 15 [14] Dettmer K, Aronov PA, Hammock BD. Mass spectrometry-based metabolomics. Mass Spectrom 416 Rev 2007;26:51-78. 417 [15] Piro JR, Benjamin DI, Duerr JM, Pi YQ, Gonzales C, Wood KM, Schwartz JW, Nomura DK, Samad 418 TA. A dysregulated endocannabinoid-eicosanoid network supports pathogenesis in a mouse model of 419 Alzheimer’s disease. Cell Rep 2012;1:617-23. 420 [16] Hu ZP, Browne ER, Liu T, Angel TE, Ho PC, Chan ECY. Metabonomic profiling of TASTPM 421 transgenic alzheimer’s disease mouse model. J Proteome Res 2012;11:5903-13. 422 [17] Trushina E, Nemutlu E, Zhang S, ChristensenT, Camp J, Mesa J, Siddiqui A, Tamura Y, Sesaki H, 423 Wengenack TM, Dzeja PP, Poduslo JF. Defects in mitochondrial dynamics and metabolomic signatures 424 of evolving energetic stress in mouse models of familial Alzheimer’s disease. PLoS ONE 2012;7:e32737. 425 [18] Wang H, Lian K, Han B, Wang Y, Kuo SH, Geng Y, Qiang J, Sun M, Wang M. Age-related 426 alterations in the metabolic profile in the hippocampus of the senescence-accelerated mouse prone 8: a 427 spontaneous Alzheimer's disease mouse model. J Alzheimers Dis 2014;39:841-8. 428 [19] Draper J, Lloyd AJ, Goodacre R, Beckmann M. Flow infusion electrospray ionisation mass 429 spectrometry for high throughput, non-targeted metabolite fingerprinting: a review. Metabolomics 430 2013;9:S4-S29. 431 [20] Lin S, Liu H, Kanawati B, Liu L, Dong J, Li M, Huang J, Schmitt-Kopplin P, Cai Z. Hippocampal 432 metabolomics using ultrahigh-resolution mass spectrometry reveals neuroinflammation from Alzheimer’s 433 disease in CRND8 mice. Anal Bioanal Chem 2013;405:5105-17. 434 [21] Lin S, Kanawati B, Liu L, Witting M, Li M, Huang J, Schmitt-Kopplin P, Cai Z. Ultra high 435 resolution mass spectrometry-based metabolic characterization reveals cerebellum as a disturbed region in 436 two animal models. Talanta 2014;118:45-53. 437 [22] Jankowsky JL, Fadale DJ, Anderson J, Xu GM, Gonzales V, Jenkins NA, Copeland NG, Lee MK, 438 Younkin LH, Wagner SL, Younkin SG, Borchelt DR. Mutant presenilins specifically elevate the levels of 439 the 42 residue beta-amyloid peptide in vivo: evidence for augmentation of a 42-specific g secretase. Hum 440 Mol Genet 2004;13:159-170. 441 [23] Sangster T, Major H, Plumb R, Wilson AJ, Wilson ID. A pragmatic and readily implemented quality 442 control strategy for HPLC-MS and GC-MS-based metabonomic analysis. Analyst 2006;131:1075-8. 443 16 [24] van den Berg RA, Hoefsloot HCJ, Westerhuis JA, Smilde AK, van der Werf MJ. Centering, scaling, 444 and transformations: improving the biological information content of metabolomics data. BMC Genomics 445 2006;7:142. 446 [25] Pulfer M, Murphy RC. Electrospray mass spectrometry of phospholipids. Mass Spectrom Rev 447 2003;22:332-64. 448 [26] Wang C, Xie S, Yang J, Yang Q, Xu G. Structural identification of human blood phospholipids using 449 liquid chromatography/quadrupole-linear ion trap mass spectrometry. Anal Chim Acta 2004;525:1-10. 450 [27] Vernez L, Hopfgartner G, Wenk M, Krahenbuhl S. Determination of carnitine and acylcarnitines in 451 urine by high-performance liquid chromatography–electrospray ionization ion trap tandem mass 452 spectrometry. J Chromatogr A 2003;984:203-13. 453 [28] Xia J, Wishart DS. MetPA: a web-based metabolomics tool for pathway analysis and visualization. 454 Bioinformatics 2010;26:2342-4. 455 [29] Lin L, Yu Q, Yan X, Hang W, Zheng J, Xing J, Huang B. Direct infusion mass spectrometry or 456 liquid chromatography mass spectrometry for human metabonomics? A serum metabonomic study of 457 kidney cancer, Analyst 2010;135:2970-8. 458 [30] Lokhov PG, Dashtiev MI, Moshkovskii SA, Archakov AI. Metabolite profiling of blood plasma of 459 patients with prostate cancer. Metabolomics 2010;6:156-63. 460 [31] Lokhov PG, Kharybin ON, Archakov AI. Diagnosis of lung cancer based on direct-infusion 461 electrospray mass spectrometry of blood plasma metabolites. Int J Mass Spectrom 2012;309:200-5. 462 [32] Gonzalez-Dominguez R, Garcia-Barrera T, Gomez-Ariza JL. Metabolomic approach to Alzheimer’s 463 disease diagnosis based on mass spectrometry. Chem Papers 2012;66:829-35. 464 [33] Gonzalez-Dominguez R, Garcia-Barrera T, Gomez-Ariza JL. Combination of metabolomic and 465 phospholipid-profiling approaches for the study of Alzheimer's disease. J Proteomics 2014;104:37-47. 466 [34] González-Domínguez R, García-Barrera T, Gómez-Ariza JL. Metabolomic study of lipids in serum 467 for biomarker discovery in Alzheimer’s disease using direct infusion mass spectrometry. J Pharm Biomed 468 Anal 2014, in press. 469 [35] Farooqui AA, Ong WY, Horrocks LA. Biochemical aspects of neurodegeneration in human brain: 470 involvement of neural membrane phospholipids and phospholipases A2. Neurochem Res 2004;29:1961-471 77. 472 17 [36] Walter A, Korth U, Hilgert M, Hartmann J, Weichel O, Hilgert M, Fassbender K, Schmitt A, Klein J. 473 Glycerophosphocholine is elevated in cerebrospinal fluid of Alzheimer patients. Neurobiol Aging 474 2004;25:1299-1303. 475 [37] Frisardi V, Panza F, Seripa D, Farooqui T, Farooqui AA. Glycerophospholipids and 476 glycerophospholipid-derived lipid mediators: A complex meshwork in Alzheimer’s disease pathology. 477 Prog Lipid Res 2011;50:313-30. 478 [38] Yao JK, Wengenack TM, Curran GL, Poduslo JF. Reduced membrane lipids in the cortex of 479 Alzheimer’s disease transgenic mice. Neurochem Res 2009;34:102-8. 480 [39] Han X, Holtzman DM, McKeel Jr DW. Plasmalogen deficiency in early Alzheimer's disease subjects 481 and in animal models: molecular characterization using electrospray ionization mass spectrometry. J 482 Neurochem 2001;77:1168-80. 483 [40] Chan RB, Oliveira TG, Cortes EP, Honig LS, Duff KE, Small SA, Wenk MR, Shui G, Di Paolo G. 484 Comparative lipidomic analysis of mouse and human brain with Alzheimer disease. J Biol Chem 485 2012;287:2678-88. 486 [41] Kou J, Kovacs GG, Höftberger R, Kulik W, Brodde A, Forss-Petter S, Hönigschnabl S, Gleiss A, 487 Brügger B, Wanders R, Just W, Budka H, Jungwirth S, Fischer P, Berger J. Peroxisomal alterations in 488 Alzheimer’s disease. Acta Neuropathol 2011;122:271-83. 489 [42] Axelsen PH, Murphy RC. Quantitative analysis of phospholipids containing arachidonate and 490 docosahexaenoate chains in microdissected regions of mouse brain. J Lipid Res 2010;51:660-71. 491 [43] Graham SF, Holscher C, Green BD. Metabolic signatures of human Alzheimer’s disease (AD): 1H 492 NMR analysis of the polar metabolome of post-mortem brain tissue. Metabolomics 2013;in press. 493 [44] Makar TK, Cooper AJ, Tofel-Grehl B, Thaler HT, Blass JP. Carnitine, carnitine acetyltransferase, 494 and glutathione in Alzheimer brain. Neurochem Res 1995;20:705-11. 495 [45] Yao J, Hamilton RT, Cadenas E, Brinton RD. Decline in mitochondrial bioenergetics and shift to 496 ketogenic profile in brain during reproductive senescence. Biochim Biophys Acta 2010;1800:1121-6. 497 [46] Yang SY, He XY, Schulz H. 3-Hydroxyacyl-CoA dehydrogenase and short chain 3-hydroxyacyl-498 CoA dehydrogenase in human health and disease. FEBS J 2005;272:4874-83. 499 [47] Cuadrado-Tejedor M, Cabodevilla JF, Zamarbide M, Gómez-Isla T, Franco R, Perez-Mediavilla A. 500 Age-related mitochondrial alterations without neuronal loss in the hippocampus of a transgenic model of 501 Alzheimer's disease. Curr Alzheimer Res 2013;10:390-405. 502 18 [48] Parnetti L, Gaiti A, Polidori MC, Brunetti M, Palumbo B, Chionne F, Cadini D, Cecchetti R, Senin 503 U. Increased cerebrospinal fluid pyruvate levels in Alzheimer’s disease. Neurosci Lett 1005;199:231-3. 504 [49] Sims B, Powers RE, Sabina RL, Theibert AB. Elevated adenosine monophosphate deaminase 505 activity in Alzheimer’s disease brain. Neurobiol Aging 1998;19:385-91. 506 [50] Cai Z, Yan LJ, Li K, Quazi SH, Zhao B. Roles of AMP-activated protein kinase in Alzheimer’s 507 disease. Neuromol Med 2012;14:1-14. 508 [51] Albasanz JL, Perez S, Barrachina M, Ferrer I, Martín M. Up-regulation of adenosine receptors in the 509 frontal cortex in Alzheimer’s disease. Brain Pathol 2008;18:211-9. 510 [52] Angulo E, Casadó V, Mallol J, Canela EI, Viñals F, Ferrer I, Lluis C, Franco R. A1 adenosine 511 receptors accumulate in neurodegenerative structures in Alzheimer’s disease and mediate both amyloid 512 precursor protein processing and tau phosphorylation and translocation. Brain Pathol 2003;13:440-51. 513 [53] Czech C, Berndt P, Busch K, Schmitz O, Wiemer J, Most V, Hampel H, Kastler J, Senn H. 514 Metabolite profiling of Alzheimer’s disease cerebrospinal fluid. PLoS ONE 2012;7:e31501. 515 [54] Lyras L, Cairns NJ, Jenner A, Jenner P, Halliwell B. An assessment of oxidative damage to proteins, 516 lipids, and DNA in brain from patients with Alzheimer's disease. J Neurochem 1997;68:2061-9. 517 [55] Gabbita SP, Lovell MA, Markesbery WR. Increased nuclear DNA oxidation in the brain in 518 Alzheimer's disease. J Neurochem 1998;71:2034-40. 519 [56] Fabelo N, Martín V, Marín R, Santpere G, Aso E, Ferrer I, Díaz M. Evidence for premature lipid raft 520 aging in APP/PS1 double-transgenic mice, a model of familial Alzheimer disease. J Neuropathol Exp 521 Neurol 2012;71:868-81. 522 [57] Strott CA, Higashi Y. Cholesterol sulfate in human physiology: what’s it all about? J Lipid Res 523 2003;44:1268-78. 524 [58] Ramalho RM, Viana RJS, Low WC, Steer CJ, Rodrigues CMP. Bile acids and apoptosis modulation: 525 an emerging role in experimental Alzheimer’s disease. Trends Mol Med 2008;14:54-62. 526 [59] Lo AC, Callaerts-Vegh Z, Nunes AF, Rodrigues CMP, D'Hooge R. Tauroursodeoxycholic acid 527 (TUDCA) supplementation prevents cognitive impairment and amyloid deposition in APP/PS1 mice. 528 Neurobiol Dis 2013;50:21-9. 529 [60] Storga D, Vrecko K, Birkmayer JGD, Reibnegger G. Monoaminergic neurotransmitters, their 530 precursors and metabolites in brains of Alzheimer patients. Neurosci Lett 1996;203:29-32. 531 19 [61] Fonteh AN, Harrington RJ, Tsai A, Liao P, Harrington MG. Free amino acid and dipeptide changes 532 in the body fluids from Alzheimer’s disease subjects. Amino Acids 2007;32:213-24. 533 [62] Samakashvili S, Ibáñez C, Simó C, Gil-Bea FJ, Winblad B, Cedazo-Mínguez A, Cifuentes A. 534 Analysis of chiral amino acids in cerebrospinal fluid samples linked to different stages of Alzheimer 535 disease. Electrophoresis 2011;32:2757-64. 536 [63] Yatin SM, Yatin M, Varadarajan S, Ain KB, Butterfield DA. Role of spermine in amyloid -peptide-537 associated free radical-induced neurotoxicity. J Neurosci Res 2001;63:395-401. 538 [64] Inoue K, Tsutsui H, Akatsu H, Hashizume Y, Matsukawa N, Yamamoto T, Toyo'oka T. Metabolic 539 profiling of Alzheimer's disease brains. Sci Rep 2013;3:2364. 540 [65] Hansmannel F, Sillaire A, Kamboh MI, Lendon C, Pasquier F, Hannequin D, Laumet G, Mounier A, 541 Ayral AM, DeKosky ST, Hauw JJ, Berr C, Mann D, Amouyel P, Campion D, Lambert JC. Is the urea 542 cycle involved in Alzheimer’s disease? J Alzheimers Dis 2010;21:1013-21. 543 [66] Perry TL, Yong VW, Bergeron C, Hansen S, Jones K. Amino acids, glutathione, and glutathione 544 transferase activity in the brains of patients with Alzheimer's disease. Ann Neurol 1987;21:331-6. 545 [67] Aoyama K, Nakaki T. Impaired glutathione synthesis in neurodegeneration. Int J Mol Sci 546 2013;14:21021-44. 547 [68] Adams JD Jr, Klaidman LK, Odunze IN, Shen HC, Miller CA. Alzheimer's and Parkinson's disease. 548 Brain levels of glutathione, glutathione disulfide, and vitamin E. Mol Chem Neuropathol. 1991;14:213-549 26. 550 551 Figure captions 552 Fig. 1. Scores plots of PCA (A) and PLS-DA (B) models for cortex. 553 Fig 2. Pathway analysis overview, where each node represents an altered metabolic pathway in brain 554 from APP/PS1 mice and its size indicates the impact of this pathway. (a) linoleic acid metabolism; (b) 555 glycerophospholipid metabolism; (c) phenylalanine, tyrosine and tryptophan metabolism; (d) glutathione 556 metabolism; (e) valine, leucine and isoleucine biosynthesis; (f) tyrosine metabolism; (g) pyruvate 557 metabolism; (h) pyrimidine metabolism; (i) glycolysis or gluconeogenesis; (j) ether lipid metabolism; (k) 558 purine metabolism. 559 Fig. 3. Overview of the most important metabolomic changes observed in brain from APP/PS1 mice. (A) 560 Metabolism of phospholipids and fatty acids. (B) Metabolism of purines and pyrimidines. Abbreviations: 561 26 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 27 Table 4. Other potential markers for discrimination between APP/PS1 and control mice. HIP, 653 hippocampus; CTX, cortex; CB, cerebellum; OB, olfactory bulbs. FC: fold change; NS: non significant 654 change. 655 Metabolite diagnostic ions HIP CTX CB OB FC (p-value) VIP FC (p-value) VIP FC (p-value) VIP FC (p-value) VIP fatty acids and related compounds palmitoleic acid N: 253.215 [M-H]- NS 1.21 (2.4·10-2) 1.77 NS NS palmitic acid N: 255.238 [M-H]- 1.18 (2.1·10-2) 1.89 1.17 (6.8·10-3) 1.85 NS NS linoleic acid N: 279.237 [M-H]- NS 1.25 (5.8·10-8) 2.56 NS 1.22 (2.3·10-3) 1.65 oleic acid N: 281.255 [M-H]- 1.18 (6.9·10-3) 1.82 1.22 (1.4·10-4) 2.31 NS NS stearic acid N: 283.268 [M-H]- 1.20 (4.0·10-4) 1.96 1.17 (2.6·10-3) 1.94 NS 1.16 (8.2·10-4) 1.70 docosahexaenoic acid N: 327.240 [M-H]- 1.22 (1.4·10-4) 2.08 1.18 (2.1·10-3) 1.78 1.26 (1.6·10-3) 2.32 NS docosapentaenoic acid N: 329.256 [M-H]- 1.23 (6.0·10-6) 2.25 1.21 (3.8·10-4) 2.19 NS NS docosatetraenoic acid N: 331.271 [M-H]- 1.41 (1.0·10-6) 3.21 1.37 (1.0·10-6) 3.26 1.35 (9.4·10-5) 2.85 1.23 (2.8·10-3) 1.97 HEPE N: 317.222 [M-H]-, 299.201, 255.212, 59.014 1.25 (9.0·10-4) 2.28 1.41 (1.2·10-5) 3.19 NS 1.56 (1.0·10-6) 3.60 HETE N: 319.228 [M-H]-, 301.209, 257.221, 59.014 1.30 (1.0·10-6) 2.67 1.33 (4.5·10-4) 2.60 1.21 (1.1·10-3) 2.18 NS nucleotide metabolism Uracil N: 111.019 [M-H]-, 42.001 1.16 (3.4·10-4) 1.68 NS NS 1.16 (6.2·10-3) 1.51 28 Adenine P: 136.063 [M+H]+, 119.039, 92.028, 65.020 N: 134.048 [M-H]-, 107.032, 92.030 0.93 (1.2·10-2) 2.45 0.93 (2.7·10-2) 1.78 0.83 (1.0·10-6) 2.60 0.87 (1.6·10-3) 1.60 hypoxanthine P: 137.045 [M+H]+, 119.035, 110.037, 94.041 N: 135.030 [M-H]-, 92.030, 65.012 NS NS 1.15 (1.0·10-6) 3.74 NS Xanthine N: 151.023 [M-H]-, 108.020, 80.028, 42.001 1.39 (1.0·10-6) 3.26 1.41 (1.0·10-6) 3.58 1.46 (2.5·10-2) 3.79 1.17 (1.4·10-2) 1.61 FAPy-adenine P: 154.068 [M+H]+, 136.061, 126.081, 119.041, 109.051 1.31 (1.8·10-3) 1.61 1.29 (1.1·10-4) 1.89 NS NS Adenosine N: 266.095 [M-H]-, 134.045, 107.041 NS NS 1.18 (1.0·10-6) 3.68 NS Inosine N: 267.082 [M-H]-, 135.030, 108.020 1.51 (1.0·10-6) 3.77 1.53 (1.0·10-6) 3.80 1.40 (1.0·10-6) 2.32 NS UMP N: 323.033 [M-H]-, 211.001, 96.970, 78.958 0.78 (1.0·10-6) 2.74 0.73 (8.6·10-4) 2.99 0.89 (5.0·10-5) 1.81 NS AMP N: 346.063 [M-H]-, 211.010, 134.045, 96.970, 78.958 0.85 (5.8·10-5) 1.83 0.84 (2.0·10-6) 2.27 NS NS steroids Cholesterol P: 369.351 [M+H-H2O]+, 287.273, 257.231, 189.162, 175.145, 161.131, 135.118, 95.082, 81.071, 57.075 0.82 (1.8·10-2) 1.72 0.74 (1.5·10-3) 2.01 NS NS deoxycholic acid P: 415.276 [M+Na]+, 357.261 0.66 (3.1·10-5) 2.28 0.66 (1.0·10-6) 1.69 NS 0.83 (1.9·10-2) 1.52 taurocholic acid P: 516.312 [M+H]+, 498.288, 480.281, 462.264, 337.251 0.81 (7.4·10-4) 2.66 0.74 (2.9·10-3) 2.02 NS NS cholesterol sulfate N: 465.304 [M-H]-, 96.061 0.74 (1.6·10-4) 2.80 0.77 (3.5·10-2) 2.26 NS NS 29 others acetic acid N: 59.014 [M-H]-, 41.012 1.19 (3.2·10-2) 1.92 NS NS NS Urea P: 61.040 [M+H]+, 44.012 0.88 (2.8·10-2) 1.60 0.69 (2.7·10-3) 2.35 NS 0.81 (3.6·10-3) 1.79 propionic acid N: 73.026 [M-H]-, 55.017 1.23 (3.4·10-2) 2.08 NS NS NS pyruvic acid N: 87.011 [M-H]-, 43.015 1.13 (2.2·10-2) 1.70 NS NS NS Alanine P: 90.055 [M+H]+, 44.050 1.19 (3.7·10-2) 1.91 NS NS NS Choline P: 104.105 [M+H]+, 60.082 1.29 (1.9·10-4) 1.98 1.21 (1.0·10-6) 2.02 1.44 (1.1·10-4) 2.76 NS Valine P: 118.084 [M+H]+, 72.082, 55.055 1.15 (1.8·10-4) 1.66 1.13 (1.1·10-3) 1.76 NS 1.18 (4.3·10-3) 1.63 aspartic acid N: 132.032 [M-H]-, 115.001, 88.041, 71.016 0.89 (1.9·10-3) 1.81 0.87 (6.0·10-6) 1.91 NS NS Dopamine N: 154.072 [M-H]-, 122.037 0.92 (1.8·10-3) 2.77 0.90 (1.8·10-4) 2.93 0.77 (2.9·10-2) 2.03 0.84 (5.5·10-4) 2.07 glycerol-3- phosphate N: 171.008 [M-H]-, 96.970, 78.958 1.19 (4.1·10-2) 1.56 1.15 (2.9·10-3) 1.52 1.16 (2.6·10-3) 1.78 NS Tyrosine P: 182.085 [M+H]+, 165.058, 147.046, 136.076, 123.048, 119.052 0.82 (3.4·10-2) 1.78 0.72 (2.4·10-2) 1.57 0.74 (7.5·10-3) 1.51 NS phosphocholine P: 184.073 [M+H]+, 104.107, 86.096 1.15 (1.7·10-3) 1.59 1.16 (2.5·10-3) 1.78 1.15 (6.9·10-3) 1.66 NS N1- acetylspermidine P: 188.176 [M+H]+, 171.152, 100.076, 72.081 NS 1.54 (1.0·10-6) 3.07 NS 1.26 (1.9·10-3) 1.95 homocarnosine P: 241.131 [M+H]+, 156.078, 110.072 0.87 (7.2·10-5) 1.70 NS NS NS GPC P: 280.091 [M+Na]+, 104.107, 1.21 1.97 1.24 1.89 NS 1.22 2.11 30 86.096 (6.2·10-5) (9.0·10-6) (4.2·10-5) Glutathione P: 308.086 [M+H]+, 233.065, 179.050, 162.025, 76.021 1.14 (1.1·10-3) 1.54 1.28 (1.0·10-6) 2.33 NS 1.24 (2.7·10-3) 1.94 GPI N: 333.061 [M-H]-, 241.021 1.21 (1.0·10-5) 2.11 1.21 (4.8·10-4) 2.00 1.19 (3.8·10-5) 2.16 1.16 (6.2·10-3) 1.63 Abbreviations: HEPE: hydroxyeicosapentaenoic acid; HETE: hydroxyeicosatetraenoic acid; UMP: 656 uridine monophosphate; AMP: adenosine monophosphate; GPC: glycerophosphocholine; GPI: 657 glycerophosphoinositol; P: positive mode; N: negative mode. 658 Figure(s) Click here to download high resolution image Figure(s) Click here to download high resolution image Figure(s) Click here to download high resolution image