- Email: [email protected]
Lipidomic profiling of plasma samples from patients with mitochondrial disease
Biochemical and Biophysical Research Communications xxx (2018) 1e8
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc
Lipidomic profiling of plasma samples from patients with mitochondrial disease Caixia Ren a, Jia Liu b, Juntuo Zhou b, Hui Liang b, Yizhang Zhu b, Qingqing Wang c, Yinglin Leng c, Zhe Zhang c, Yun Yuan c, Zhaoxia Wang c, **, Yuxin Yin b, d, * a
Departments of Human Anatomy, Histology and Embryology, Peking University Health Science Center, Beijing 100191, China Institute of Systems Biomedicine, Department of Pathology, School of Basic Medical Sciences, Peking University Health Science Center, Beijing 100191, China c Department of Neurology, Peking University First Hospital, Beijing 100034, China d Beijing Key Laboratory of Tumor Systems Biology, Peking-Tsinghua Center for Life Sciences, Beijing 100191, China b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 15 March 2018 Accepted 20 March 2018 Available online xxx
Mitochondrial disease (MD) is a rare mitochondrial respiratory chain disorder with a high mortality and extremely challenging to treat. Although genomic, transcriptomic, and proteomic analyses have been performed to investigate the pathogenesis of MD, the role of metabolomics in MD, particularly of lipidomics remains unclear. This study was undertaken to identify potential lipid biomarkers of MD. An untargeted lipidomic approach was used to compare the plasma lipid metabolites in 20 MD patients and 20 controls through Liquid Chromatography coupled to Mass Spectrometry. Volcano plot analysis was performed to identify the different metabolites. Receiver operating characteristic (ROC) curves were constructed and the area under the ROC curves (AUC) was calculated to determine the potentially sensitive and specific biomarkers. A total of 41 lipids were significantly different in MD patients and controls. ROC curve analysis showed the top 5 AUC values of lipids (phosphatidylinositols 38:6, lysoPC 20:0, 19:0, 18:0, 17:0) are more than 0.99. Multivariate ROC curve based exploratory analysis showed the AUC of combination of top 5 lipids is 1, indicating they may be potentially sensitive and specific biomarkers for MD. We propose combination of these lipid species may be more valuable in predicting the development and progression of MD, and this will have important implications for the diagnosis and treatment of MD. © 2018 Elsevier Inc. All rights reserved.
Keywords: Mitochondrial disease Lipidomics Biomarkers
1. Introduction Mitochondrial disease (MD) refers to a rare primary mitochondrial disorder that results in inadequate energy production. The most often affected tissues and organs are those with the highest energy demands, for example skeletal muscle, brain or heart. MD is a result of mutations either in the nuclear or mitochondrial DNA (mtDNA) [1]. The lowest prevalence of MD in adults is ~12.5 per 100 000, and ~4.7 per 100 000 in children [2]. Clinical symptoms can arise in childhood or later in life, and can affect one organ in
* Corresponding author. Institute of Systems Biomedicine, Department of Pathology, School of Basic Medical Sciences, Peking University Health Science Center, Beijing 100191, China. ** Corresponding author. E-mail addresses: [email protected] (Z. Wang), [email protected] (Y. Yin).
isolation or present as multisystem disease [3]. Many geneticallycharacterized mitochondrial disorders present with a combination of clinical features which are characteristic of distinct MD syndromes, such as mitochondrial encephalomyopathy with lactic acidosis and stroke like episodes (MELAS) [4], Kearns-Sayre syndrome (KSS) [5], chronic progressive external ophthalmoplegia (CPEO) [6], mitochondrial limb girdle myopathy (MLGM), isolated mitochondrial myopathy (MM) [7], myoclonic epilepsy with ragged red fibres (MERRF) [8] and so on. Diagnosing MD is a challenge because of its genetic heterogeneity, diversity of clinical phenotypes [9]. In addition, significant difficulties may also be encountered in the management of MD. Despite progress in current understanding of the pathophysiology and genetics of MD, no effective cure for mitochondrial disorders has been found [10]. Apart from supportive therapy, a variety of therapeutic approaches have been evaluated in randomized clinical trials, but unfortunately none of these has delivered breakthrough results [11,12]. Therefore,
https://doi.org/10.1016/j.bbrc.2018.03.160 0006-291X/© 2018 Elsevier Inc. All rights reserved.
Please cite this article in press as: C. Ren, et al., Lipidomic profiling of plasma samples from patients with mitochondrial disease, Biochemical and Biophysical Research Communications (2018), https://doi.org/10.1016/j.bbrc.2018.03.160
2
C. Ren et al. / Biochemical and Biophysical Research Communications xxx (2018) 1e8
increased understanding of MD characteristics is needed to facilitate diagnosis and treatment of MD. Lipids are crucial components of cellular membranes and lipid particles such as lipoproteins. Lipids play many essential roles in cellular functions, including cellular barriers, membrane matrices, signaling, and energy depots [13]. Lipidomics has led to identification of new signaling molecules, motivated discovery of potential biomarkers for early diagnosis and treatment of disease, supported screening of drug targets and/or test drug efficiency, and allowed personalization of medical treatment [14]. So far there is no report on the lipidomics study in MD patients yet. Lipid droplet accumulation has been observed in the skeletal muscle of patients with MD with transmission electron microscopy [15] and magnetic resonance spectroscopy (MRS) [16]. This led us to suppose altered lipid metabolism occurs in MD. In this study, we aimed to examine plasma lipidomic metabolites with Liquid Chromatography coupled with Mass Spectrometry (LC-MS) to identify potential MD biomarkers, which may provide fundamental information which is of benefit for diagnosis and treatment of MD.
study patients. Diagnosis of MD was undertaken based on the clinical phenotype, muscle pathology, and genetic criteria. None of these patients were taking any medication except for treatment of MD. All subjects gave informed consent for inclusion before participation in this study. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Ethics Committee of the Peking University First Hospital (2012542) (Supplementary material 1).
2. Materials and methods
Lipids were extracted from plasma samples with a modified Folch method [17]. Typically, 100 mL of plasma were aliquoted into a 0.6 ml Eppendorf tube and mixed with 400 mL of chloroform/ methanol (2:1, V/V) containing 20 mg/ml of free fatty acid 19:0 as an internal standard. After vortexing for 10 min, the mixture was centrifuged at 13000 rpm at 4 C for 20 min. The lower lipid containing chloroform phase was evaporated with a speed vacuum, and the residue was stored at 80 C for further analysis. All samples were processed in the same laboratory to avoid bias.
2.1. Patients Twenty patients with MD (10 MELAS, 3 KSS, 6 CPEO and 1 MERRF) and 20 healthy control subjects were recruited for untargeted lipidomic analysis (Table S1). The control subjects in whom MD was excluded by examination were recruited from hospital and laboratory personnel, and were age and sex matched with the
2.2. Chemicals and reagents Formic acid, HPLC grade methanol, acetonitrile (ACN) and isopropanol (IPA) were obtained from Fisher Scientific. Chloroform was obtained from Tong Guang Fine Chemicals Company (Beijing, China). Ultra-pure water was supplied by a Millipore system (Millipore, Billerica, MA, USA). 2.3. Sample preparations for untargeted lipidomic profiling
Fig. 1. PCA (A) and orthoPLS-DA (B) score plots of MD patients based on metabolomics data in positive ion mode. PCA (D) and orthoPLS-DA (E) score plots of MD patients based on metabolomics data in negative ion mode. Performance statistics of orthoPLS-DA in positive ion mode (C) and negative ion mode (F).
Please cite this article in press as: C. Ren, et al., Lipidomic profiling of plasma samples from patients with mitochondrial disease, Biochemical and Biophysical Research Communications (2018), https://doi.org/10.1016/j.bbrc.2018.03.160
C. Ren et al. / Biochemical and Biophysical Research Communications xxx (2018) 1e8
3
Fig. 2. Heat map of 41 altered metabolites in MD compared to controls. Red, upregulated; blue, downregulated. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
2.4. High-performance liquid chromatography and mass spectrometry for untargeted lipidomic profiling An Ultimate 3000 ultra high performance liquid chromatography (UHPLC) system coupled to Q-Exactive MS (Thermo Scientific) was used for lipid separation and detection. Chromatographic separation was performed on a reversed phase X select CSH C18 column (4.6 mm 100 mm, 2.5 mm, Waters, USA) consistent with what has been reported [17]. (Supplementary material 2).
based exploratory analysis was performed to explore the performance of different biomarker models. (MetaboAnalyst:www. metaboanalyst.ca). 2.6. Statistical analysis Age, height, weight and BMI and Boxplots analysis were carried out using SPSS statistical software (version 16.0). The differences between groups were compared using t tests. Significance was set at p < 0.05.
2.5. Data processing and analysis for untargeted lipidomic profiling 3. Results The acquired raw data were processed using MSDIAL according to the instructions in the software tutorial [18]. Datasets containing m/z values, retention time, and peak area were exported as an Excel file, and the Excel file was imported into the Metabo Analyst 3.0 Web service [19] for multivariate analysis. Principal component analysis (PCA) which is an unsupervised chemometric method, was used to obtain an overall picture of the data sets in their entirety, and determine whether there was any clustering, trends, or outliers. Orthogonal partial least squares-discriminant analysis (orthoPLS-DA) was used for further data analysis. Volcano plot analysis with fold-change (FC) > 1.5 and false discovery rate (FDR) < 0.05 by Student's t-test was performed to identify the differential metabolites. Heat map was generated with R. Receiver operating characteristic (ROC) curves were constructed and the area under the ROC curves (AUC) was calculated to investigate whether the metabolites could be efficiently exploited for constructing a sensitive biomarker of MD. Multivariate ROC curve
3.1. Identification of potential metabolic biomarkers PCA and orthogonal partial least squares-discriminant analysis (orthoPLS-DA) score plots including MD patients, healthy controls and quality control (QC) samples are shown in Fig. 1. QC samples (blue) clustered together tightly in positive ion mode (Fig. 1A) and negative ion mode (Fig. 1D), reflecting the stability of the LC-MS system and showing that the quality of all the LC-MS data for this study were satisfactory. The MD group (M group, green) and healthy control group (C group, pink) were clustered together and separated from each other in positive ion mode (Fig. 1B) and negative ion mode (Fig. 1E). This suggested that the involved metabolites were perturbed in the MD group. The percentage of extracted variance related to class information was 14.2% in positive ion mode (Fig. 1C) and 13.6% in negative ion mode (Fig. 1F) and explained 82.9% of the variance of class variables in positive ion
Please cite this article in press as: C. Ren, et al., Lipidomic profiling of plasma samples from patients with mitochondrial disease, Biochemical and Biophysical Research Communications (2018), https://doi.org/10.1016/j.bbrc.2018.03.160
4
C. Ren et al. / Biochemical and Biophysical Research Communications xxx (2018) 1e8
Fig. 3. Boxplots representing relative abundance of: (A) lysoPC 22:5, (B) lysoPC 24:1, (C) lysoPC 22:1, (D) lysoPC 16:0, (E) lysoPC 16:1, (F) lysoPC 20:2, (G) lysoPC 18:1, (H) lysoPC 14:0, (I) lysoPC17:0, (J) lysoPC20:0, (K) lysoPC19:0, (L) lysoPC18:0, (M) lysoPC15:0, (N) lysoPC20:1 in positive ion mode.
Please cite this article in press as: C. Ren, et al., Lipidomic profiling of plasma samples from patients with mitochondrial disease, Biochemical and Biophysical Research Communications (2018), https://doi.org/10.1016/j.bbrc.2018.03.160
C. Ren et al. / Biochemical and Biophysical Research Communications xxx (2018) 1e8
5
Table 1 AUC0.90, 95% confidence intervals (95% CI), and sensitivity and specificity for ROC curves, together with p-values for 20 significant lipidomic metabolites. No.
Significant Metabolites
AUC
95% CI
Sensitivity
Specificity
p value
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
PI 38:6 lysoPC 20:0 lysoPC 19:0 lysoPC 18:0 lysoPC 17:0 PC 24:0 lysoPC 20:2 lysoPE 18:0 LysoPC 24:1 lysoPC 16:0 lysoPC 15:0 lysoPC 14:0 lysoPC 18:1 lysoPC 22:1 PE36:4 PE34:1 Acylcarnitine Acylcarnitine Acylcarnitine Acylcarnitine
1 1 1 0.992 0.99 0.988 0.972 0.968 0.955 0.95 0.938 0.935 0.938 0.932 0.92 0.912 0.941 0.935 0.92 0.912
1e1 0.985e1 0.985e1 0.96e1 0.955e1 0.952e1 0.916e1 0.903e1 0.871e1 0.845e1 0.821e1 0.819e1 0.842e0.994 0.821e1 0.815e0.98 0.819e0.975 0.844e1 0.829e1 0.83e0.981 0.796e0.99
1 1 1 1 0.9 1 1 1 0.9 1 0.9 0.8 0.8 0.9 0.8 0.8 1 0.9 0.9 0.9
1 1 1 1 1 1 1 0.9 0.9 1 1 1 0.9 1 0.9 0.9 0.9 0.9 0.8 0.8
0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
12:2 16:0 20:3 18:0
mode (R2Y ¼ 0.829) and 80.7% of the variance of class variables in negative ion mode (R2Y ¼ 0.807). These multivariate analyses showed significant metabolic differences in the MD and control groups. A total of 2137 identified metabolites in positive ion mode and 714 metabolites in negative ion mode were evaluated in these two groups. After quality control analysis and further volcano plot analysis with FC > 1.5 and FDR< 0.05 using the Student's t-test, 33 lipidomic metabolites in positive ion mode and 8 in negative ion mode were significantly different in MD patients and controls, including lysophosphatidylcholines (lysoPCs), phosphatidylcholines (PCs), lysophosphatidylethanolamines (lysoPEs), phosphatidylethanolamines (PEs), sphingomyelins (SMs); acylcarnitines, triglycerides (TGs) and phosphatidyl inositols (PI) (Table S2). The heat map showed these 41 altered metabolites both in positive ion mode and negative ion mode (Fig. 2). Of note, 14 lysoPCs were dramatically decreased (Fig. 3), and PIs (38:6, 38:4, 36:2), PEs (36:4, 34:1), acylcarnitines (11:1, 20:3, 12:2, 18:0, 16:0), TGs (52:1, 50:0) unsaturated PCs (PC 32:1, 36:5, 34:3) and unsaturated lysoPEs (lysoPE 20:3, 22:4, 18:2) were significantly increased in MD patients compared to controls. Moreover, PE 36:4, PE 34:1, lysoPE 20:3 were detected in both positive ion mode and negative ion mode. 3.2. Diagnostic performance of metabolites ROC curve analysis of potential biomarker levels for differentiating MD patients was performed. Of these 41 metabolites, 20 showed AUC >0.90 when MD patients and controls were compared (Table 1). Lipids with AUC0.95 are shown in Fig. 4 (AeI). The top 5 AUC values of the lipids (PI 38:6, lysoPC 20:0 and lysoPC 19:0, lysoPC 18:0, lysoPC 17:0) were more than 0.99 (Table 1, Fig. 4). Multivariate ROC curve based exploratory analysis showed the AUC of combination of the top 5 lipids is 1 (Fig. 4J and K), indicating these lipids may be potentially sensitive and specific biomarkers for MD. 4. Discussion The mitochondrion is found in all cells except the mature erythrocyte. It is the source of the majority of energy production and free radical production, and is responsible for calcium homeostasis, apoptosis, innate immunity, and inflammation [20]. This
is the first study to our knowledge to utilize a lipidomic platform with LC-MS to examine lipid species in plasma of MD. We identified 41 metabolites that were significantly different in MD patients and controls, and identified potentially sensitive and specific biomarkers of MD.
4.1. Elevated acylcarnitine and TG in MD Acylcarnitines are intermediates in fatty acid oxidation [21], and are synthesized by carnitine palmitoyltransferase 1 (CPT1) which serves in the transport of fatty acids into the mitochondrial matrix. It has now become clear that a net efflux of acylcarnitine species from the mitochondria into the cytosol and ultimately into plasma is particularly important under conditions of impaired fatty acid oxidation for prevention of accumulation of potentially toxic acylCoA intermediates in the mitochondrion [22]. Incomplete fatty acid oxidation results in elevated concentrations of acylcarnitine [23]. In this study, MD patients exhibited increased levels of medium-to long-chain acylcarnitines, indicating impairment of fatty acid oxidation is involved in MD which may result from dysfunction of CPT1. Excessive metabolic energy is stored as TG in lipid droplets in adipocytes. Intramyocellular lipid accumulation was observed in MD patients [15,16]. Elevated acylcarnitines and TG suggest decreased fuel utilization and increased deposit of TG occur in MD patients.
4.2. Increased PI in MD PI plays important roles in cell signaling and in the regulation of membrane traffic and transport functions [24]. It is a precursor of phosphatidylinositol 4,5-bisphosphate (PIP2), which is catalyzed by phosphatidylinositol phospholipid-specific phospholipase C (PLC) into IP3 which is a regulator of cytosolic calcium levels, and diacylglycerol (DAG), which is an activator of protein kinase C [25]. PIs (38:6, 38:4, 36:2) that are elevated in MD may be derived from dysfunction of PLC and consequently affect calcium levels. The AUC value of PI 38:6 was 1, suggesting PI 38:6 may be a potentially sensitive and specific biomarker for MD.
Please cite this article in press as: C. Ren, et al., Lipidomic profiling of plasma samples from patients with mitochondrial disease, Biochemical and Biophysical Research Communications (2018), https://doi.org/10.1016/j.bbrc.2018.03.160
6
C. Ren et al. / Biochemical and Biophysical Research Communications xxx (2018) 1e8
Fig. 4. ROC curves for (A) PI 38:6; (B) lysoPC 20:0; (C) lysoPC 19:0; (D) lysoPC 18:0; (E) lysoPC 17:0; (F) PC; 24:0; (G) lysoPC 20:2; (H) lysoPE 18:0; (I) lysoPC 16:0 using AUC0.95 for comparing MD and control groups. Multivariate ROC curve based exploratory analysis curves for the top 5 lipids (PI 38:6, lysoPC 20:0, lysoPC 19:0, lysoPC 18:0, lysoPC 17:0) in (J) comparing all models and (K) combination analysis of 5 lipids in MD and control groups.
Please cite this article in press as: C. Ren, et al., Lipidomic profiling of plasma samples from patients with mitochondrial disease, Biochemical and Biophysical Research Communications (2018), https://doi.org/10.1016/j.bbrc.2018.03.160
C. Ren et al. / Biochemical and Biophysical Research Communications xxx (2018) 1e8
4.3. Up-regulated PE and unsatuated lysoPE in MD PE is a minor constituent of cell membranes, and plays important roles in biological processes such as apoptosis and cell signaling [26]. PE-binding probes have been shown to be selective in detecting dead and dying cells [27]. LysoPE is a metabolic product of PE which is generated by phospholipase A2 [28]. LysoPE has been shown to mobilize intracellular Ca2þ through G-proteincoupled receptor (GPCR) in some cells types [29e31]. In fibroblasts from patients with MD, levels of ionized Ca2þ at rest are elevated compared to controls [32]. Elevated lysoPEs (20:3, 22:4, 18:2) may lead to altered calcium homeostasis in MD patients, and subsequently result in cell death. 4.4. Elevated PC and decreased lysoPCs in MD PC represents the main membrane-forming phospholipid in mammalian cells. PE undergoes three successive methylation reactions by PE N-methyl transferase (PEMT) in its full conversion to PC [33]. LysoPC is derived from PC, and is produced mainly by two pathways. The first is the result of partial hydrolysis of PC, in which one of the fatty acids is removed by the action of phospholipase A2 (PLA2) [34]. A second pathway for lysoPC formation occurs by transfer of one fatty acid of PC to cholesterol by lecithin-cholesterol acyltransferase (LCAT), which is an enzyme that converts free cholesterol into cholesteryl ester [35]. Unsaturated PC species (32:1, 36:5, 34:3) were increased in MD, and may have been generated from elevated unsatuated PEs. Fourteen LysoPCs were decreased in MD patients, and the AUC values of 11 lysoPCs were more than 0.9. Among the top 5 lipids, 4 lysoPCs (lysoPC 20:0,19:0, 18:0,17:0) had high AUC values (more than 0.99). Multivariate ROC curve based exploratory analysis revealed these lysoPCs may be potentially sensitive and specific biomarkers for MD. The reason lysoPCs depletion in MD is unclear. Dysfunction of PLA2 or LCAT may lead to low levels of lysoPCs production and accumulation of PCs in MD. Decreased lysoPCs have been found in cancer and inflammatory disorders, and lysoPC concentration was found to correlate negatively with C-reactive protein (CRP) [36e38]. Administration of lysoPC protected against lethality of sepsis accompanied by decreased TNFa and IL-1b levels [39,40]. LysoPC effects on inflammatory processes may be mediated, at least in part by lysoPC binding G protein coupled receptor (GPR) G2A, which appears to play a functional role in the modulation of activation, migration and apoptosis of a variety of immune cells including neutrophils, macrophages and lymphocytes [41] and the elimination of dead eukaryotic and prokaryotic cells during infection [38]. Depletion of plasma lysoPC has also been observed in patients with Alzheimers Disease and mild cognitive impairment [42]. Management of PLA2 or LCAT and/or exogenous supplementation of these lysoPCs may be beneficial for treatment of MD. In summary, we determined 41 lipidomic metabolites were significantly different in MD and controls, and 5 of those were potentially sensitive and specific biomarkers for MD. Lipids have diverse biological roles and their metabolism is extremely complex. It is unlikely that a single lipid species would be suitable as an MD biomarker. Combinations of several lipid species may be of value in predicting the development and progression of MD and diagnosis of MD. We propose targeted manipulation of these increased lipids and lipid related enzymes or exogenous supplementation of decreased lysoPCs may be of benefit in the treatment of MD. Further studies in larger populations and additional cohorts are warranted to validate the conclusions of this study. More comprehensive studies combining genomics, transcriptomics, proteomics, and metabolomics should be conducted to describe the detailed
7
variations in MD and to support precise steps of prevention and intervention. Acknowledgements This work was supported by grants to Yuxin Yin from National Key Research and Development Program of China (Grant #2016YFA0500302), National Natural Science Foundation of China (Key grants #81430056, #31420103905 and #81621063), Beijing Natural Science Foundation (Key grant #7161007), and Lam Chung Nin Foundation for Systems Biomedicine; Grant to Zhaoxia Wang from the Beijing Municipal Science and Technology Commission (No. Z151100003915126), and grants to Caixia Ren from the National Natural Science Foundation of China (81402388); Leading Academic Discipline Project of Beijing Education Bureau (BMU20110254) and Peking University grant (BMU20150492). Appendix A. Supplementary data Supplementary data related to this article can be found at https://doi.org/10.1016/j.bbrc.2018.03.160. Transparency document Transparency document related to this article can be found online at https://doi.org/10.1016/j.bbrc.2018.03.160. References [1] W.J. Koopman, P.H. Willems, J.A. Smeitink, Monogenic mitochondrial disorders, N. Engl. J. Med. 366 (2012) 1132e1141. [2] D. Skladal, J. Halliday, D.R. Thorburn, Minimum birth prevalence of mitochondrial respiratory chain disorders in children, Brain 126 (2003) 1905e1912. [3] C.L. Alston, M.C. Rocha, N.Z. Lax, D.M. Turnbull, R.W. Taylor, The genetics and pathology of mitochondrial disease, J. Pathol. 241 (2017) 236e250. [4] A.W. El-Hattab, A.M. Adesina, J. Jones, F. Scaglia, Melas syndrome: clinical manifestations, pathogenesis, and treatment options, Mol. Genet. Metabol. 116 (2015) 4e12. [5] F.J. Carod-Artal, E. Lopez Gallardo, A. Solano, Y. Dahmani, M.D. Herrero, J. Montoya, [mitochondrial DNA deletions in kearns-sayre syndrome], Neurologia 21 (2006) 357e364. [6] C. Richardson, T. Smith, A. Schaefer, D. Turnbull, P. Griffiths, Ocular motility findings in chronic progressive external ophthalmoplegia, Eye (Lond) 19 (2005) 258e263. [7] Yuanyuan Lu, DZ, Sheng Yao, Shiwen Wu, Daojun Hong, Qingqing Wang, Jing Liua, Jan A.M. Smeitink, Yun Yuan, Zhaoxia Wang, Mitochondrial trna genes are hotspots for mutations in a cohort of patients with exercise intolerance and mitochondrial myopathy, J. Neurol. Sci. 379 (2017) 137e143. [8] T. Tsutsumi, H. Nishida, Y. Noguchi, A. Komatsuzaki, K. Kitamura, Audiological findings in patients with myoclonic epilepsy associated with ragged-red fibres, J. Laryngol. Otol. 115 (2001) 777e781. [9] J. Finsterer, Mitochondriopathies, Eur. J. Neurol. 11 (2004) 163e186. [10] J.A. Smeitink, M. Zeviani, D.M. Turnbull, H.T. Jacobs, Mitochondrial medicine: a metabolic perspective on the pathology of oxidative phosphorylation disorders, Cell Metabol. 3 (2006) 9e13. [11] G. Pfeffer, K. Majamaa, D.M. Turnbull, D. Thorburn, P.F. Chinnery, Treatment for mitochondrial disorders, Cochrane Database Syst. Rev. (2012) CD004426. [12] M. Kanabus, S.J. Heales, S. Rahman, Development of pharmacological strategies for mitochondrial disorders, Br. J. Pharmacol. 171 (2014) 1798e1817. [13] L. Yetukuri, M. Katajamaa, G. Medina-Gomez, T. Seppanen-Laakso, A. VidalPuig, M. Oresic, Bioinformatics strategies for lipidomics analysis: characterization of obesity related hepatic steatosis, BMC Syst. Biol. 1 (2007) 12. [14] K. Yang, X. Han, Lipidomics: techniques, applications, and outcomes related to biomedical sciences, Trends Biochem. Sci. 41 (2016) 954e969. [15] J.Y. Mun, M.K. Jung, S.H. Kim, S. Eom, S.S. Han, Y.M. Lee, Ultrastructural changes in skeletal muscle of infants with mitochondrial respiratory chain complex i defects, J. Clin. Neurol. 13 (2017) 359e365. [16] S. Golla, J. Ren, C.R. Malloy, J.M. Pascual, Intramyocellular lipid excess in the mitochondrial disorder melas: mrs determination at 7t, Neurol. Genet. 3 (2017) e160. [17] C. Ren, J. Liu, J. Zhou, H. Liang, Y. Wang, Y. Sun, B. Ma, Y. Yin, Lipidomic analysis of serum samples from migraine patients, Lipids Health Dis. 17 (2018) 22. [18] H. Tsugawa, T. Cajka, T. Kind, Y. Ma, B. Higgins, K. Ikeda, M. Kanazawa, J. VanderGheynst, O. Fiehn, M. Arita, Ms-dial: data-independent ms/ms deconvolution for comprehensive metabolome analysis, Nat. Meth. 12 (2015)
Please cite this article in press as: C. Ren, et al., Lipidomic profiling of plasma samples from patients with mitochondrial disease, Biochemical and Biophysical Research Communications (2018), https://doi.org/10.1016/j.bbrc.2018.03.160
8
C. Ren et al. / Biochemical and Biophysical Research Communications xxx (2018) 1e8
523e526. [19] J. Xia, N. Psychogios, N. Young, D.S. Wishart, Metaboanalyst: a web server for metabolomic data analysis and interpretation, Nucleic Acids Res. 37 (2009) W652eW660. [20] B.H. Cohen, Neuromuscular and systemic presentations in adults: diagnoses beyond merrf and melas, Neurotherapeutics 10 (2013) 227e242. [21] R.R. Ramsay, R.D. Gandour, F.R. van der Leij, Molecular enzymology of carnitine transfer and transport, Biochim. Biophys. Acta 1546 (2001) 21e43. [22] S. Violante, L. Ijlst, J. Ruiter, J. Koster, H. van Lenthe, M. Duran, I.T. de Almeida, R.J. Wanders, S.M. Houten, F.V. Ventura, Substrate specificity of human carnitine acetyltransferase: implications for fatty acid and branched-chain amino acid metabolism, Biochim. Biophys. Acta 1832 (2013) 773e779. [23] P. Giesbertz, J. Ecker, A. Haag, B. Spanier, H. Daniel, An lc-ms/ms method to quantify acylcarnitine species including isomeric and odd-numbered forms in plasma and tissues, J. Lipid Res. 56 (2015) 2029e2039. [24] G. Di Paolo, P. De Camilli, Phosphoinositides in cell regulation and membrane dynamics, Nature 443 (2006) 651e657. [25] M.J. Rebecchi, S.N. Pentyala, Structure, function, and control of phosphoinositide-specific phospholipase c, Physiol. Rev. 80 (2000) 1291e1335. [26] L.L.M. Xue, T. Chen, H. Sun, J. Zhu, X. Li, F. Wu, B. Wang, J. Li, Y. Chen, Peinduced apoptosis in smmc-7721 cells: involvement of erk and stat signalling pathways, Int. J. Mol. Med. 34 (2014) 119e129. [27] M. Zhao, Z. Li, S. Bugenhagen, 99mtc-labeled duramycin as a novel phosphatidylethanolamine-binding molecular probe, J. Nucl. Med. 49 (2008) 1345e1352. [28] S.K. Lau, LK, G.C. Lo, V.S. Ding, W.N. Chow, T.Y. Ke, S.O. Curreem, K.K. To, D.T. Ho, S. Sridhar, S.C. Wong, J.F. Chan, I.F. Hung, K.H. Sze, C.W. Lam, K.Y. Yuen, P.C. Woo, Metabolomic profiling of plasma from melioidosis patients using uhplc-qtof ms reveals novel biomarkers for diagnosis, Int. J. Mol. Sci. 17 (2016) 307. [29] S.J. Park, K.P. Lee, D.S. Im, Action and signaling of lysophosphatidylethanolamine in mda-mb-231 breast cancer cells, Biomol. Therapeut. (Seoul) 22 (2014) 129e135. [30] K.S. Park, H.Y. Lee, S.Y. Lee, M.K. Kim, S.D. Kim, J.M. Kim, J. Yun, D.S. Im, Y.S. Bae, Lysophosphatidylethanolamine stimulates chemotactic migration and cellular invasion in sk-ov3 human ovarian cancer cells: involvement of pertussis toxin-sensitive g-protein coupled receptor, FEBS Lett. 581 (2007) 4411e4416. [31] J.M.P.S. Lee, D.S. Im, Calcium signaling of lysophosphatidylethanolamine through lpa1 in human sh-sy5y neuroblastoma cells, Biomol. Therapeut.
(Seoul) 25 (2017) 194e201. [32] A.M. Moudy, S.D. Handran, M.P. Goldberg, N. Ruffin, I. Karl, P. Kranz-Eble, D.C. DeVivo, S.M. Rothman, Abnormal calcium homeostasis and mitochondrial polarization in a human encephalomyopathy, Proc. Natl. Acad. Sci. U. S. A. 92 (1995) 729e733. [33] E. Calzada, O. Onguka, S.M. Claypool, Phosphatidylethanolamine metabolism in health and disease, Int. Rev. Cell Mol. Biol. 321 (2016) 29e88. [34] M.H.J.A. Cha, M.M. Ko, C. Zhang, M.S. Lee, Metabolic profiles distinguish nondampness-phlegm and dampness-phlegm patterns among Korean patients with acute cerebral infarction, Evid. Base. Compl. Alternative Med. 2013 (2013) 9. Article ID 517018. [35] T. Matsumoto, T. Kobayashi, K. Kamata, Role of lysophosphatidylcholine (lpc) in atherosclerosis, Curr. Med. Chem. 14 (2007) 3209e3220. [36] L.A.A.J. Taylor, A.K. Hodina, C. Unger, U. Massing, Plasma lysophosphatidylcholine concentration is decreased in cancer patients with weight loss and activated inflammatory status, Lipids Health Dis. 6 (2007) 1e8. ski, Application of 31p nmr spectroscopy [37] M.J.W. Kuliszkiewicz-Janus, S. Baczyn in clinical analysis of changes of serum phospholipids in leukemia, lymphoma and some other non-haematological cancers, Anticancer Res. 16 (1996) 1587e1594. [38] R. Parra Millan, M.E. Jimenez Mejias, V. Sanchez Encinales, R. Ayerbe Algaba, A. Gutierrez Valencia, M.E. Pachon Ibanez, C. Diaz, J. Perez Del Palacio, L.F. Lopez Cortes, J. Pachon, Y. Smani, Efficacy of lysophosphatidylcholine in combination with antimicrobial agents against acinetobacter baumannii in experimental murine peritoneal sepsis and pneumonia models, Antimicrob. Agents Chemother. 60 (2016) 4464e4470. [39] J.J. Yan, J.S. Jung, J.E. Lee, J. Lee, S.O. Huh, H.S. Kim, K.C. Jung, J.Y. Cho, J.S. Nam, H.W. Suh, Y.H. Kim, D.K. Song, Therapeutic effects of lysophosphatidylcholine in experimental sepsis, Nat. Med. 10 (2004) 161e167. [40] W. Drobnik, G. Liebisch, F.X. Audebert, D. Frohlich, T. Gluck, P. Vogel, G. Rothe, G. Schmitz, Plasma ceramide and lysophosphatidylcholine inversely correlate with mortality in sepsis patients, J. Lipid Res. 44 (2003) 754e761. [41] J.H. Kabarowski, G2a and lpc: regulatory functions in immunity, Prostag. Other Lipid Mediat. 89 (2009) 73e81. [42] M. Mapstone, A.K. Cheema, M.S. Fiandaca, X. Zhong, T.R. Mhyre, L.H. MacArthur, W.J. Hall, S.G. Fisher, D.R. Peterson, J.M. Haley, M.D. Nazar, S.A. Rich, D.J. Berlau, C.B. Peltz, M.T. Tan, C.H. Kawas, H.J. Federoff, Plasma phospholipids identify antecedent memory impairment in older adults, Nat. Med. 20 (2014) 415e418.
Please cite this article in press as: C. Ren, et al., Lipidomic profiling of plasma samples from patients with mitochondrial disease, Biochemical and Biophysical Research Communications (2018), https://doi.org/10.1016/j.bbrc.2018.03.160
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc
Lipidomic profiling of plasma samples from patients with mitochondrial disease Caixia Ren a, Jia Liu b, Juntuo Zhou b, Hui Liang b, Yizhang Zhu b, Qingqing Wang c, Yinglin Leng c, Zhe Zhang c, Yun Yuan c, Zhaoxia Wang c, **, Yuxin Yin b, d, * a
Departments of Human Anatomy, Histology and Embryology, Peking University Health Science Center, Beijing 100191, China Institute of Systems Biomedicine, Department of Pathology, School of Basic Medical Sciences, Peking University Health Science Center, Beijing 100191, China c Department of Neurology, Peking University First Hospital, Beijing 100034, China d Beijing Key Laboratory of Tumor Systems Biology, Peking-Tsinghua Center for Life Sciences, Beijing 100191, China b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 15 March 2018 Accepted 20 March 2018 Available online xxx
Mitochondrial disease (MD) is a rare mitochondrial respiratory chain disorder with a high mortality and extremely challenging to treat. Although genomic, transcriptomic, and proteomic analyses have been performed to investigate the pathogenesis of MD, the role of metabolomics in MD, particularly of lipidomics remains unclear. This study was undertaken to identify potential lipid biomarkers of MD. An untargeted lipidomic approach was used to compare the plasma lipid metabolites in 20 MD patients and 20 controls through Liquid Chromatography coupled to Mass Spectrometry. Volcano plot analysis was performed to identify the different metabolites. Receiver operating characteristic (ROC) curves were constructed and the area under the ROC curves (AUC) was calculated to determine the potentially sensitive and specific biomarkers. A total of 41 lipids were significantly different in MD patients and controls. ROC curve analysis showed the top 5 AUC values of lipids (phosphatidylinositols 38:6, lysoPC 20:0, 19:0, 18:0, 17:0) are more than 0.99. Multivariate ROC curve based exploratory analysis showed the AUC of combination of top 5 lipids is 1, indicating they may be potentially sensitive and specific biomarkers for MD. We propose combination of these lipid species may be more valuable in predicting the development and progression of MD, and this will have important implications for the diagnosis and treatment of MD. © 2018 Elsevier Inc. All rights reserved.
Keywords: Mitochondrial disease Lipidomics Biomarkers
1. Introduction Mitochondrial disease (MD) refers to a rare primary mitochondrial disorder that results in inadequate energy production. The most often affected tissues and organs are those with the highest energy demands, for example skeletal muscle, brain or heart. MD is a result of mutations either in the nuclear or mitochondrial DNA (mtDNA) [1]. The lowest prevalence of MD in adults is ~12.5 per 100 000, and ~4.7 per 100 000 in children [2]. Clinical symptoms can arise in childhood or later in life, and can affect one organ in
* Corresponding author. Institute of Systems Biomedicine, Department of Pathology, School of Basic Medical Sciences, Peking University Health Science Center, Beijing 100191, China. ** Corresponding author. E-mail addresses: [email protected] (Z. Wang), [email protected] (Y. Yin).
isolation or present as multisystem disease [3]. Many geneticallycharacterized mitochondrial disorders present with a combination of clinical features which are characteristic of distinct MD syndromes, such as mitochondrial encephalomyopathy with lactic acidosis and stroke like episodes (MELAS) [4], Kearns-Sayre syndrome (KSS) [5], chronic progressive external ophthalmoplegia (CPEO) [6], mitochondrial limb girdle myopathy (MLGM), isolated mitochondrial myopathy (MM) [7], myoclonic epilepsy with ragged red fibres (MERRF) [8] and so on. Diagnosing MD is a challenge because of its genetic heterogeneity, diversity of clinical phenotypes [9]. In addition, significant difficulties may also be encountered in the management of MD. Despite progress in current understanding of the pathophysiology and genetics of MD, no effective cure for mitochondrial disorders has been found [10]. Apart from supportive therapy, a variety of therapeutic approaches have been evaluated in randomized clinical trials, but unfortunately none of these has delivered breakthrough results [11,12]. Therefore,
https://doi.org/10.1016/j.bbrc.2018.03.160 0006-291X/© 2018 Elsevier Inc. All rights reserved.
Please cite this article in press as: C. Ren, et al., Lipidomic profiling of plasma samples from patients with mitochondrial disease, Biochemical and Biophysical Research Communications (2018), https://doi.org/10.1016/j.bbrc.2018.03.160
2
C. Ren et al. / Biochemical and Biophysical Research Communications xxx (2018) 1e8
increased understanding of MD characteristics is needed to facilitate diagnosis and treatment of MD. Lipids are crucial components of cellular membranes and lipid particles such as lipoproteins. Lipids play many essential roles in cellular functions, including cellular barriers, membrane matrices, signaling, and energy depots [13]. Lipidomics has led to identification of new signaling molecules, motivated discovery of potential biomarkers for early diagnosis and treatment of disease, supported screening of drug targets and/or test drug efficiency, and allowed personalization of medical treatment [14]. So far there is no report on the lipidomics study in MD patients yet. Lipid droplet accumulation has been observed in the skeletal muscle of patients with MD with transmission electron microscopy [15] and magnetic resonance spectroscopy (MRS) [16]. This led us to suppose altered lipid metabolism occurs in MD. In this study, we aimed to examine plasma lipidomic metabolites with Liquid Chromatography coupled with Mass Spectrometry (LC-MS) to identify potential MD biomarkers, which may provide fundamental information which is of benefit for diagnosis and treatment of MD.
study patients. Diagnosis of MD was undertaken based on the clinical phenotype, muscle pathology, and genetic criteria. None of these patients were taking any medication except for treatment of MD. All subjects gave informed consent for inclusion before participation in this study. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Ethics Committee of the Peking University First Hospital (2012542) (Supplementary material 1).
2. Materials and methods
Lipids were extracted from plasma samples with a modified Folch method [17]. Typically, 100 mL of plasma were aliquoted into a 0.6 ml Eppendorf tube and mixed with 400 mL of chloroform/ methanol (2:1, V/V) containing 20 mg/ml of free fatty acid 19:0 as an internal standard. After vortexing for 10 min, the mixture was centrifuged at 13000 rpm at 4 C for 20 min. The lower lipid containing chloroform phase was evaporated with a speed vacuum, and the residue was stored at 80 C for further analysis. All samples were processed in the same laboratory to avoid bias.
2.1. Patients Twenty patients with MD (10 MELAS, 3 KSS, 6 CPEO and 1 MERRF) and 20 healthy control subjects were recruited for untargeted lipidomic analysis (Table S1). The control subjects in whom MD was excluded by examination were recruited from hospital and laboratory personnel, and were age and sex matched with the
2.2. Chemicals and reagents Formic acid, HPLC grade methanol, acetonitrile (ACN) and isopropanol (IPA) were obtained from Fisher Scientific. Chloroform was obtained from Tong Guang Fine Chemicals Company (Beijing, China). Ultra-pure water was supplied by a Millipore system (Millipore, Billerica, MA, USA). 2.3. Sample preparations for untargeted lipidomic profiling
Fig. 1. PCA (A) and orthoPLS-DA (B) score plots of MD patients based on metabolomics data in positive ion mode. PCA (D) and orthoPLS-DA (E) score plots of MD patients based on metabolomics data in negative ion mode. Performance statistics of orthoPLS-DA in positive ion mode (C) and negative ion mode (F).
Please cite this article in press as: C. Ren, et al., Lipidomic profiling of plasma samples from patients with mitochondrial disease, Biochemical and Biophysical Research Communications (2018), https://doi.org/10.1016/j.bbrc.2018.03.160
C. Ren et al. / Biochemical and Biophysical Research Communications xxx (2018) 1e8
3
Fig. 2. Heat map of 41 altered metabolites in MD compared to controls. Red, upregulated; blue, downregulated. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
2.4. High-performance liquid chromatography and mass spectrometry for untargeted lipidomic profiling An Ultimate 3000 ultra high performance liquid chromatography (UHPLC) system coupled to Q-Exactive MS (Thermo Scientific) was used for lipid separation and detection. Chromatographic separation was performed on a reversed phase X select CSH C18 column (4.6 mm 100 mm, 2.5 mm, Waters, USA) consistent with what has been reported [17]. (Supplementary material 2).
based exploratory analysis was performed to explore the performance of different biomarker models. (MetaboAnalyst:www. metaboanalyst.ca). 2.6. Statistical analysis Age, height, weight and BMI and Boxplots analysis were carried out using SPSS statistical software (version 16.0). The differences between groups were compared using t tests. Significance was set at p < 0.05.
2.5. Data processing and analysis for untargeted lipidomic profiling 3. Results The acquired raw data were processed using MSDIAL according to the instructions in the software tutorial [18]. Datasets containing m/z values, retention time, and peak area were exported as an Excel file, and the Excel file was imported into the Metabo Analyst 3.0 Web service [19] for multivariate analysis. Principal component analysis (PCA) which is an unsupervised chemometric method, was used to obtain an overall picture of the data sets in their entirety, and determine whether there was any clustering, trends, or outliers. Orthogonal partial least squares-discriminant analysis (orthoPLS-DA) was used for further data analysis. Volcano plot analysis with fold-change (FC) > 1.5 and false discovery rate (FDR) < 0.05 by Student's t-test was performed to identify the differential metabolites. Heat map was generated with R. Receiver operating characteristic (ROC) curves were constructed and the area under the ROC curves (AUC) was calculated to investigate whether the metabolites could be efficiently exploited for constructing a sensitive biomarker of MD. Multivariate ROC curve
3.1. Identification of potential metabolic biomarkers PCA and orthogonal partial least squares-discriminant analysis (orthoPLS-DA) score plots including MD patients, healthy controls and quality control (QC) samples are shown in Fig. 1. QC samples (blue) clustered together tightly in positive ion mode (Fig. 1A) and negative ion mode (Fig. 1D), reflecting the stability of the LC-MS system and showing that the quality of all the LC-MS data for this study were satisfactory. The MD group (M group, green) and healthy control group (C group, pink) were clustered together and separated from each other in positive ion mode (Fig. 1B) and negative ion mode (Fig. 1E). This suggested that the involved metabolites were perturbed in the MD group. The percentage of extracted variance related to class information was 14.2% in positive ion mode (Fig. 1C) and 13.6% in negative ion mode (Fig. 1F) and explained 82.9% of the variance of class variables in positive ion
Please cite this article in press as: C. Ren, et al., Lipidomic profiling of plasma samples from patients with mitochondrial disease, Biochemical and Biophysical Research Communications (2018), https://doi.org/10.1016/j.bbrc.2018.03.160
4
C. Ren et al. / Biochemical and Biophysical Research Communications xxx (2018) 1e8
Fig. 3. Boxplots representing relative abundance of: (A) lysoPC 22:5, (B) lysoPC 24:1, (C) lysoPC 22:1, (D) lysoPC 16:0, (E) lysoPC 16:1, (F) lysoPC 20:2, (G) lysoPC 18:1, (H) lysoPC 14:0, (I) lysoPC17:0, (J) lysoPC20:0, (K) lysoPC19:0, (L) lysoPC18:0, (M) lysoPC15:0, (N) lysoPC20:1 in positive ion mode.
Please cite this article in press as: C. Ren, et al., Lipidomic profiling of plasma samples from patients with mitochondrial disease, Biochemical and Biophysical Research Communications (2018), https://doi.org/10.1016/j.bbrc.2018.03.160
C. Ren et al. / Biochemical and Biophysical Research Communications xxx (2018) 1e8
5
Table 1 AUC0.90, 95% confidence intervals (95% CI), and sensitivity and specificity for ROC curves, together with p-values for 20 significant lipidomic metabolites. No.
Significant Metabolites
AUC
95% CI
Sensitivity
Specificity
p value
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
PI 38:6 lysoPC 20:0 lysoPC 19:0 lysoPC 18:0 lysoPC 17:0 PC 24:0 lysoPC 20:2 lysoPE 18:0 LysoPC 24:1 lysoPC 16:0 lysoPC 15:0 lysoPC 14:0 lysoPC 18:1 lysoPC 22:1 PE36:4 PE34:1 Acylcarnitine Acylcarnitine Acylcarnitine Acylcarnitine
1 1 1 0.992 0.99 0.988 0.972 0.968 0.955 0.95 0.938 0.935 0.938 0.932 0.92 0.912 0.941 0.935 0.92 0.912
1e1 0.985e1 0.985e1 0.96e1 0.955e1 0.952e1 0.916e1 0.903e1 0.871e1 0.845e1 0.821e1 0.819e1 0.842e0.994 0.821e1 0.815e0.98 0.819e0.975 0.844e1 0.829e1 0.83e0.981 0.796e0.99
1 1 1 1 0.9 1 1 1 0.9 1 0.9 0.8 0.8 0.9 0.8 0.8 1 0.9 0.9 0.9
1 1 1 1 1 1 1 0.9 0.9 1 1 1 0.9 1 0.9 0.9 0.9 0.9 0.8 0.8
0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
12:2 16:0 20:3 18:0
mode (R2Y ¼ 0.829) and 80.7% of the variance of class variables in negative ion mode (R2Y ¼ 0.807). These multivariate analyses showed significant metabolic differences in the MD and control groups. A total of 2137 identified metabolites in positive ion mode and 714 metabolites in negative ion mode were evaluated in these two groups. After quality control analysis and further volcano plot analysis with FC > 1.5 and FDR< 0.05 using the Student's t-test, 33 lipidomic metabolites in positive ion mode and 8 in negative ion mode were significantly different in MD patients and controls, including lysophosphatidylcholines (lysoPCs), phosphatidylcholines (PCs), lysophosphatidylethanolamines (lysoPEs), phosphatidylethanolamines (PEs), sphingomyelins (SMs); acylcarnitines, triglycerides (TGs) and phosphatidyl inositols (PI) (Table S2). The heat map showed these 41 altered metabolites both in positive ion mode and negative ion mode (Fig. 2). Of note, 14 lysoPCs were dramatically decreased (Fig. 3), and PIs (38:6, 38:4, 36:2), PEs (36:4, 34:1), acylcarnitines (11:1, 20:3, 12:2, 18:0, 16:0), TGs (52:1, 50:0) unsaturated PCs (PC 32:1, 36:5, 34:3) and unsaturated lysoPEs (lysoPE 20:3, 22:4, 18:2) were significantly increased in MD patients compared to controls. Moreover, PE 36:4, PE 34:1, lysoPE 20:3 were detected in both positive ion mode and negative ion mode. 3.2. Diagnostic performance of metabolites ROC curve analysis of potential biomarker levels for differentiating MD patients was performed. Of these 41 metabolites, 20 showed AUC >0.90 when MD patients and controls were compared (Table 1). Lipids with AUC0.95 are shown in Fig. 4 (AeI). The top 5 AUC values of the lipids (PI 38:6, lysoPC 20:0 and lysoPC 19:0, lysoPC 18:0, lysoPC 17:0) were more than 0.99 (Table 1, Fig. 4). Multivariate ROC curve based exploratory analysis showed the AUC of combination of the top 5 lipids is 1 (Fig. 4J and K), indicating these lipids may be potentially sensitive and specific biomarkers for MD. 4. Discussion The mitochondrion is found in all cells except the mature erythrocyte. It is the source of the majority of energy production and free radical production, and is responsible for calcium homeostasis, apoptosis, innate immunity, and inflammation [20]. This
is the first study to our knowledge to utilize a lipidomic platform with LC-MS to examine lipid species in plasma of MD. We identified 41 metabolites that were significantly different in MD patients and controls, and identified potentially sensitive and specific biomarkers of MD.
4.1. Elevated acylcarnitine and TG in MD Acylcarnitines are intermediates in fatty acid oxidation [21], and are synthesized by carnitine palmitoyltransferase 1 (CPT1) which serves in the transport of fatty acids into the mitochondrial matrix. It has now become clear that a net efflux of acylcarnitine species from the mitochondria into the cytosol and ultimately into plasma is particularly important under conditions of impaired fatty acid oxidation for prevention of accumulation of potentially toxic acylCoA intermediates in the mitochondrion [22]. Incomplete fatty acid oxidation results in elevated concentrations of acylcarnitine [23]. In this study, MD patients exhibited increased levels of medium-to long-chain acylcarnitines, indicating impairment of fatty acid oxidation is involved in MD which may result from dysfunction of CPT1. Excessive metabolic energy is stored as TG in lipid droplets in adipocytes. Intramyocellular lipid accumulation was observed in MD patients [15,16]. Elevated acylcarnitines and TG suggest decreased fuel utilization and increased deposit of TG occur in MD patients.
4.2. Increased PI in MD PI plays important roles in cell signaling and in the regulation of membrane traffic and transport functions [24]. It is a precursor of phosphatidylinositol 4,5-bisphosphate (PIP2), which is catalyzed by phosphatidylinositol phospholipid-specific phospholipase C (PLC) into IP3 which is a regulator of cytosolic calcium levels, and diacylglycerol (DAG), which is an activator of protein kinase C [25]. PIs (38:6, 38:4, 36:2) that are elevated in MD may be derived from dysfunction of PLC and consequently affect calcium levels. The AUC value of PI 38:6 was 1, suggesting PI 38:6 may be a potentially sensitive and specific biomarker for MD.
Please cite this article in press as: C. Ren, et al., Lipidomic profiling of plasma samples from patients with mitochondrial disease, Biochemical and Biophysical Research Communications (2018), https://doi.org/10.1016/j.bbrc.2018.03.160
6
C. Ren et al. / Biochemical and Biophysical Research Communications xxx (2018) 1e8
Fig. 4. ROC curves for (A) PI 38:6; (B) lysoPC 20:0; (C) lysoPC 19:0; (D) lysoPC 18:0; (E) lysoPC 17:0; (F) PC; 24:0; (G) lysoPC 20:2; (H) lysoPE 18:0; (I) lysoPC 16:0 using AUC0.95 for comparing MD and control groups. Multivariate ROC curve based exploratory analysis curves for the top 5 lipids (PI 38:6, lysoPC 20:0, lysoPC 19:0, lysoPC 18:0, lysoPC 17:0) in (J) comparing all models and (K) combination analysis of 5 lipids in MD and control groups.
Please cite this article in press as: C. Ren, et al., Lipidomic profiling of plasma samples from patients with mitochondrial disease, Biochemical and Biophysical Research Communications (2018), https://doi.org/10.1016/j.bbrc.2018.03.160
C. Ren et al. / Biochemical and Biophysical Research Communications xxx (2018) 1e8
4.3. Up-regulated PE and unsatuated lysoPE in MD PE is a minor constituent of cell membranes, and plays important roles in biological processes such as apoptosis and cell signaling [26]. PE-binding probes have been shown to be selective in detecting dead and dying cells [27]. LysoPE is a metabolic product of PE which is generated by phospholipase A2 [28]. LysoPE has been shown to mobilize intracellular Ca2þ through G-proteincoupled receptor (GPCR) in some cells types [29e31]. In fibroblasts from patients with MD, levels of ionized Ca2þ at rest are elevated compared to controls [32]. Elevated lysoPEs (20:3, 22:4, 18:2) may lead to altered calcium homeostasis in MD patients, and subsequently result in cell death. 4.4. Elevated PC and decreased lysoPCs in MD PC represents the main membrane-forming phospholipid in mammalian cells. PE undergoes three successive methylation reactions by PE N-methyl transferase (PEMT) in its full conversion to PC [33]. LysoPC is derived from PC, and is produced mainly by two pathways. The first is the result of partial hydrolysis of PC, in which one of the fatty acids is removed by the action of phospholipase A2 (PLA2) [34]. A second pathway for lysoPC formation occurs by transfer of one fatty acid of PC to cholesterol by lecithin-cholesterol acyltransferase (LCAT), which is an enzyme that converts free cholesterol into cholesteryl ester [35]. Unsaturated PC species (32:1, 36:5, 34:3) were increased in MD, and may have been generated from elevated unsatuated PEs. Fourteen LysoPCs were decreased in MD patients, and the AUC values of 11 lysoPCs were more than 0.9. Among the top 5 lipids, 4 lysoPCs (lysoPC 20:0,19:0, 18:0,17:0) had high AUC values (more than 0.99). Multivariate ROC curve based exploratory analysis revealed these lysoPCs may be potentially sensitive and specific biomarkers for MD. The reason lysoPCs depletion in MD is unclear. Dysfunction of PLA2 or LCAT may lead to low levels of lysoPCs production and accumulation of PCs in MD. Decreased lysoPCs have been found in cancer and inflammatory disorders, and lysoPC concentration was found to correlate negatively with C-reactive protein (CRP) [36e38]. Administration of lysoPC protected against lethality of sepsis accompanied by decreased TNFa and IL-1b levels [39,40]. LysoPC effects on inflammatory processes may be mediated, at least in part by lysoPC binding G protein coupled receptor (GPR) G2A, which appears to play a functional role in the modulation of activation, migration and apoptosis of a variety of immune cells including neutrophils, macrophages and lymphocytes [41] and the elimination of dead eukaryotic and prokaryotic cells during infection [38]. Depletion of plasma lysoPC has also been observed in patients with Alzheimers Disease and mild cognitive impairment [42]. Management of PLA2 or LCAT and/or exogenous supplementation of these lysoPCs may be beneficial for treatment of MD. In summary, we determined 41 lipidomic metabolites were significantly different in MD and controls, and 5 of those were potentially sensitive and specific biomarkers for MD. Lipids have diverse biological roles and their metabolism is extremely complex. It is unlikely that a single lipid species would be suitable as an MD biomarker. Combinations of several lipid species may be of value in predicting the development and progression of MD and diagnosis of MD. We propose targeted manipulation of these increased lipids and lipid related enzymes or exogenous supplementation of decreased lysoPCs may be of benefit in the treatment of MD. Further studies in larger populations and additional cohorts are warranted to validate the conclusions of this study. More comprehensive studies combining genomics, transcriptomics, proteomics, and metabolomics should be conducted to describe the detailed
7
variations in MD and to support precise steps of prevention and intervention. Acknowledgements This work was supported by grants to Yuxin Yin from National Key Research and Development Program of China (Grant #2016YFA0500302), National Natural Science Foundation of China (Key grants #81430056, #31420103905 and #81621063), Beijing Natural Science Foundation (Key grant #7161007), and Lam Chung Nin Foundation for Systems Biomedicine; Grant to Zhaoxia Wang from the Beijing Municipal Science and Technology Commission (No. Z151100003915126), and grants to Caixia Ren from the National Natural Science Foundation of China (81402388); Leading Academic Discipline Project of Beijing Education Bureau (BMU20110254) and Peking University grant (BMU20150492). Appendix A. Supplementary data Supplementary data related to this article can be found at https://doi.org/10.1016/j.bbrc.2018.03.160. Transparency document Transparency document related to this article can be found online at https://doi.org/10.1016/j.bbrc.2018.03.160. References [1] W.J. Koopman, P.H. Willems, J.A. Smeitink, Monogenic mitochondrial disorders, N. Engl. J. Med. 366 (2012) 1132e1141. [2] D. Skladal, J. Halliday, D.R. Thorburn, Minimum birth prevalence of mitochondrial respiratory chain disorders in children, Brain 126 (2003) 1905e1912. [3] C.L. Alston, M.C. Rocha, N.Z. Lax, D.M. Turnbull, R.W. Taylor, The genetics and pathology of mitochondrial disease, J. Pathol. 241 (2017) 236e250. [4] A.W. El-Hattab, A.M. Adesina, J. Jones, F. Scaglia, Melas syndrome: clinical manifestations, pathogenesis, and treatment options, Mol. Genet. Metabol. 116 (2015) 4e12. [5] F.J. Carod-Artal, E. Lopez Gallardo, A. Solano, Y. Dahmani, M.D. Herrero, J. Montoya, [mitochondrial DNA deletions in kearns-sayre syndrome], Neurologia 21 (2006) 357e364. [6] C. Richardson, T. Smith, A. Schaefer, D. Turnbull, P. Griffiths, Ocular motility findings in chronic progressive external ophthalmoplegia, Eye (Lond) 19 (2005) 258e263. [7] Yuanyuan Lu, DZ, Sheng Yao, Shiwen Wu, Daojun Hong, Qingqing Wang, Jing Liua, Jan A.M. Smeitink, Yun Yuan, Zhaoxia Wang, Mitochondrial trna genes are hotspots for mutations in a cohort of patients with exercise intolerance and mitochondrial myopathy, J. Neurol. Sci. 379 (2017) 137e143. [8] T. Tsutsumi, H. Nishida, Y. Noguchi, A. Komatsuzaki, K. Kitamura, Audiological findings in patients with myoclonic epilepsy associated with ragged-red fibres, J. Laryngol. Otol. 115 (2001) 777e781. [9] J. Finsterer, Mitochondriopathies, Eur. J. Neurol. 11 (2004) 163e186. [10] J.A. Smeitink, M. Zeviani, D.M. Turnbull, H.T. Jacobs, Mitochondrial medicine: a metabolic perspective on the pathology of oxidative phosphorylation disorders, Cell Metabol. 3 (2006) 9e13. [11] G. Pfeffer, K. Majamaa, D.M. Turnbull, D. Thorburn, P.F. Chinnery, Treatment for mitochondrial disorders, Cochrane Database Syst. Rev. (2012) CD004426. [12] M. Kanabus, S.J. Heales, S. Rahman, Development of pharmacological strategies for mitochondrial disorders, Br. J. Pharmacol. 171 (2014) 1798e1817. [13] L. Yetukuri, M. Katajamaa, G. Medina-Gomez, T. Seppanen-Laakso, A. VidalPuig, M. Oresic, Bioinformatics strategies for lipidomics analysis: characterization of obesity related hepatic steatosis, BMC Syst. Biol. 1 (2007) 12. [14] K. Yang, X. Han, Lipidomics: techniques, applications, and outcomes related to biomedical sciences, Trends Biochem. Sci. 41 (2016) 954e969. [15] J.Y. Mun, M.K. Jung, S.H. Kim, S. Eom, S.S. Han, Y.M. Lee, Ultrastructural changes in skeletal muscle of infants with mitochondrial respiratory chain complex i defects, J. Clin. Neurol. 13 (2017) 359e365. [16] S. Golla, J. Ren, C.R. Malloy, J.M. Pascual, Intramyocellular lipid excess in the mitochondrial disorder melas: mrs determination at 7t, Neurol. Genet. 3 (2017) e160. [17] C. Ren, J. Liu, J. Zhou, H. Liang, Y. Wang, Y. Sun, B. Ma, Y. Yin, Lipidomic analysis of serum samples from migraine patients, Lipids Health Dis. 17 (2018) 22. [18] H. Tsugawa, T. Cajka, T. Kind, Y. Ma, B. Higgins, K. Ikeda, M. Kanazawa, J. VanderGheynst, O. Fiehn, M. Arita, Ms-dial: data-independent ms/ms deconvolution for comprehensive metabolome analysis, Nat. Meth. 12 (2015)
Please cite this article in press as: C. Ren, et al., Lipidomic profiling of plasma samples from patients with mitochondrial disease, Biochemical and Biophysical Research Communications (2018), https://doi.org/10.1016/j.bbrc.2018.03.160
8
C. Ren et al. / Biochemical and Biophysical Research Communications xxx (2018) 1e8
523e526. [19] J. Xia, N. Psychogios, N. Young, D.S. Wishart, Metaboanalyst: a web server for metabolomic data analysis and interpretation, Nucleic Acids Res. 37 (2009) W652eW660. [20] B.H. Cohen, Neuromuscular and systemic presentations in adults: diagnoses beyond merrf and melas, Neurotherapeutics 10 (2013) 227e242. [21] R.R. Ramsay, R.D. Gandour, F.R. van der Leij, Molecular enzymology of carnitine transfer and transport, Biochim. Biophys. Acta 1546 (2001) 21e43. [22] S. Violante, L. Ijlst, J. Ruiter, J. Koster, H. van Lenthe, M. Duran, I.T. de Almeida, R.J. Wanders, S.M. Houten, F.V. Ventura, Substrate specificity of human carnitine acetyltransferase: implications for fatty acid and branched-chain amino acid metabolism, Biochim. Biophys. Acta 1832 (2013) 773e779. [23] P. Giesbertz, J. Ecker, A. Haag, B. Spanier, H. Daniel, An lc-ms/ms method to quantify acylcarnitine species including isomeric and odd-numbered forms in plasma and tissues, J. Lipid Res. 56 (2015) 2029e2039. [24] G. Di Paolo, P. De Camilli, Phosphoinositides in cell regulation and membrane dynamics, Nature 443 (2006) 651e657. [25] M.J. Rebecchi, S.N. Pentyala, Structure, function, and control of phosphoinositide-specific phospholipase c, Physiol. Rev. 80 (2000) 1291e1335. [26] L.L.M. Xue, T. Chen, H. Sun, J. Zhu, X. Li, F. Wu, B. Wang, J. Li, Y. Chen, Peinduced apoptosis in smmc-7721 cells: involvement of erk and stat signalling pathways, Int. J. Mol. Med. 34 (2014) 119e129. [27] M. Zhao, Z. Li, S. Bugenhagen, 99mtc-labeled duramycin as a novel phosphatidylethanolamine-binding molecular probe, J. Nucl. Med. 49 (2008) 1345e1352. [28] S.K. Lau, LK, G.C. Lo, V.S. Ding, W.N. Chow, T.Y. Ke, S.O. Curreem, K.K. To, D.T. Ho, S. Sridhar, S.C. Wong, J.F. Chan, I.F. Hung, K.H. Sze, C.W. Lam, K.Y. Yuen, P.C. Woo, Metabolomic profiling of plasma from melioidosis patients using uhplc-qtof ms reveals novel biomarkers for diagnosis, Int. J. Mol. Sci. 17 (2016) 307. [29] S.J. Park, K.P. Lee, D.S. Im, Action and signaling of lysophosphatidylethanolamine in mda-mb-231 breast cancer cells, Biomol. Therapeut. (Seoul) 22 (2014) 129e135. [30] K.S. Park, H.Y. Lee, S.Y. Lee, M.K. Kim, S.D. Kim, J.M. Kim, J. Yun, D.S. Im, Y.S. Bae, Lysophosphatidylethanolamine stimulates chemotactic migration and cellular invasion in sk-ov3 human ovarian cancer cells: involvement of pertussis toxin-sensitive g-protein coupled receptor, FEBS Lett. 581 (2007) 4411e4416. [31] J.M.P.S. Lee, D.S. Im, Calcium signaling of lysophosphatidylethanolamine through lpa1 in human sh-sy5y neuroblastoma cells, Biomol. Therapeut.
(Seoul) 25 (2017) 194e201. [32] A.M. Moudy, S.D. Handran, M.P. Goldberg, N. Ruffin, I. Karl, P. Kranz-Eble, D.C. DeVivo, S.M. Rothman, Abnormal calcium homeostasis and mitochondrial polarization in a human encephalomyopathy, Proc. Natl. Acad. Sci. U. S. A. 92 (1995) 729e733. [33] E. Calzada, O. Onguka, S.M. Claypool, Phosphatidylethanolamine metabolism in health and disease, Int. Rev. Cell Mol. Biol. 321 (2016) 29e88. [34] M.H.J.A. Cha, M.M. Ko, C. Zhang, M.S. Lee, Metabolic profiles distinguish nondampness-phlegm and dampness-phlegm patterns among Korean patients with acute cerebral infarction, Evid. Base. Compl. Alternative Med. 2013 (2013) 9. Article ID 517018. [35] T. Matsumoto, T. Kobayashi, K. Kamata, Role of lysophosphatidylcholine (lpc) in atherosclerosis, Curr. Med. Chem. 14 (2007) 3209e3220. [36] L.A.A.J. Taylor, A.K. Hodina, C. Unger, U. Massing, Plasma lysophosphatidylcholine concentration is decreased in cancer patients with weight loss and activated inflammatory status, Lipids Health Dis. 6 (2007) 1e8. ski, Application of 31p nmr spectroscopy [37] M.J.W. Kuliszkiewicz-Janus, S. Baczyn in clinical analysis of changes of serum phospholipids in leukemia, lymphoma and some other non-haematological cancers, Anticancer Res. 16 (1996) 1587e1594. [38] R. Parra Millan, M.E. Jimenez Mejias, V. Sanchez Encinales, R. Ayerbe Algaba, A. Gutierrez Valencia, M.E. Pachon Ibanez, C. Diaz, J. Perez Del Palacio, L.F. Lopez Cortes, J. Pachon, Y. Smani, Efficacy of lysophosphatidylcholine in combination with antimicrobial agents against acinetobacter baumannii in experimental murine peritoneal sepsis and pneumonia models, Antimicrob. Agents Chemother. 60 (2016) 4464e4470. [39] J.J. Yan, J.S. Jung, J.E. Lee, J. Lee, S.O. Huh, H.S. Kim, K.C. Jung, J.Y. Cho, J.S. Nam, H.W. Suh, Y.H. Kim, D.K. Song, Therapeutic effects of lysophosphatidylcholine in experimental sepsis, Nat. Med. 10 (2004) 161e167. [40] W. Drobnik, G. Liebisch, F.X. Audebert, D. Frohlich, T. Gluck, P. Vogel, G. Rothe, G. Schmitz, Plasma ceramide and lysophosphatidylcholine inversely correlate with mortality in sepsis patients, J. Lipid Res. 44 (2003) 754e761. [41] J.H. Kabarowski, G2a and lpc: regulatory functions in immunity, Prostag. Other Lipid Mediat. 89 (2009) 73e81. [42] M. Mapstone, A.K. Cheema, M.S. Fiandaca, X. Zhong, T.R. Mhyre, L.H. MacArthur, W.J. Hall, S.G. Fisher, D.R. Peterson, J.M. Haley, M.D. Nazar, S.A. Rich, D.J. Berlau, C.B. Peltz, M.T. Tan, C.H. Kawas, H.J. Federoff, Plasma phospholipids identify antecedent memory impairment in older adults, Nat. Med. 20 (2014) 415e418.
Please cite this article in press as: C. Ren, et al., Lipidomic profiling of plasma samples from patients with mitochondrial disease, Biochemical and Biophysical Research Communications (2018), https://doi.org/10.1016/j.bbrc.2018.03.160