MS ... - ACS Publications

Subscriber access provided by LUNDS UNIV

Article

A mathematical model-assisted UHPLC-MS/MS method for global profiling and quantification of cholesteryl esters in hyperlipidemic golden hamsters Miao Lin, Zhe Wang, Dongmei Wang, Xiong Chen, and Jinlan Zhang Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b05337 • Publication Date (Web): 06 Mar 2019 Downloaded from http://pubs.acs.org on March 6, 2019

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

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

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

Analytical Chemistry

A mathematical model-assisted UHPLC-MS/MS method for global profiling and quantification of cholesteryl esters in hyperlipidemic golden hamsters Miao Lin #, Zhe Wang #, Dongmei Wang, Xiong Chen, Jin-Lan Zhang * State Key Laboratory of Bioactive Substances and Functions of Natural Medicines, Institute of Materia Medica, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100050, P. R. China. ABSTRACT: Cholesteryl esters (CEs) are formed by the 3-hydroxyl group of cholesterol and a fatty acyl chain through an ester bond, and function as a biologically inert storage form of cholesterol. Abnormal CE levels are often related to various diseases, particularly hyperlipidemia and atherosclerosis. Herein, we developed a mathematical model-assisted ultra-high performance liquid chromatography-mass spectrometry (UHPLC-MS) method for the untargeted identification to targeted quantification of CEs in plasma, different density lipoprotein samples from humans, rats, and golden hamsters. Using UHPLC-quadrupole-time-of-flight mass spectrometry (UHPLC-QTOF-MS), 81 CE candidates were detected in the above samples, of which 24 CEs were reported in the Human Metabolome Database and 57 CEs were newly identified based on an in-house database of theoretically possible CEs, including the computationally generated precursor ion m/z mass, carbon number and double bond numbers of the fatty acyl chain. Then three mathematical models based on the characteristic chromatographic retention behavior related to structural features were established and validated using commercial and synthetic CE standards. The mathematical model-assisted UHPLC-MS/MS strategy was proposed to provide a global profile and identification of CEs, especially unknown CEs. With the efficient strategy, 74 CEs in plasma of golden hamster were identified, and then quantified in normal and hyperlipidemic golden hamsters by dynamic multiple reaction monitoring (dMRM). 21 CEs among 35 shared potential biomarkers were newly found for hyperlipidemia. Our study will contribute to the in-depth study the functions of CEs and the discovery of biomarkers for diseases.

Introduction Cholesteryl esters (CEs) function as a biologically inert storage form of cholesterol (Chol) in all types of cells [1]. In vivo, cholesterol homeostasis is essential [2], as it maintains the balance between biologically active free Chol and inactive CEs. In cholesterol metabolism, Chol is usually transferred to hydrophobic CEs, which then form lipoprotein particles for lipid transportation by the circulation [3]. The maintenance of cholesterol homeostasis mainly depends on the synthesis and hydrolysis of CEs, which are biosynthesized by two different pathways. One occurs inside cells [4, 5] via acyl-CoA cholesterol acyltransferase (ACAT) to esterify cholesterol. The other occurs in extracellular spaces such as the bloodstream, where high density lipoproteins (HDL) reverse cholesterol transport catalyzed by lecithin cholesterol acyl transferase (LCAT) [6]. The hydrolysis of CEs requires cholesterol ester hydrolases [7], especially neutral cholesterol ester hydrolase (NCEH) and lysosomal acidic lipase (LAL). LAL is the primary enzyme responsible for the hydrolysis of cholesteryl esters in low density lipoproteins (LDL) [7]. Given the importance of cholesterol homeostasis in the normal state, abnormal levels of CEs are often related to various pathological conditions, such as Wolman disease [8], hyperlipidemia [9], atherosclerosis [10], Alzheimer's disease [11], and cancer [12, 13]. CE (14:0), CE (16:0), and CE (18:0) in plasma were reported as potential biomarkers for hyperlipidemia [14]. CE (18:0), CE (18:1), and CE (18:3) were potential biomarkers for hypercholesterolemia [15]. And Padró et al [16] found that

hypercholesterolemia significantly elevated the levels of CEs in the high density lipoproteins (HDL) of pigs, especially CE (16:1), CE (18:1), CE (18:0), CE (18:3), and CE (17:0). However, few studies have focused on the global profiling of CEs in plasma, VLDL, LDL, and HDL simultaneously. Cholesteryl esters are a combination of the 3-hydroxyl group of cholesterol and fatty acyl chain (FA) formed through an ester bond. Twenty-four CEs without C = C location isomers have been reported in the Human Metabolome Database (HMDB). Currently, fatty acids are reported to consist of approximately 100 compounds in HMDB. In view of the variety of fatty acids and two biosynthetic pathways of CEs, there might be many unknown CEs with different lengths, degrees of unsaturation (n, the number of C=C of FA), and C=C location isomers of the fatty acyl chain, even oxidized fatty acyl chain. Given the importance of CEs in physiological and pathological situations, research efforts in the past decades have focused on the reliable identification and quantification of CEs. To date, methods including fluorescence [17] (cholesterol assay kits), gas chromatography-mass spectrometry (GC-MS) [18], liquid chromatography-mass spectrometry (LC-MS) [19], and nuclear magnetic resonance spectroscopy (NMR) [20] have been used for CEs analysis. In clinical testing, the fluorescence method is often used for quantification of the total content of CEs and Chol. But the diversity of CEs cannot be determined using this method. Son et al [18] used GC-MS to quantify 6 CEs quantification (0.2 μg/ml) in serum from patients with vasospastic angina. However, polyunsaturated CEs are unstable and degrade easily under high-temperature condition. Therefore,

ACS Paragon Plus Environment

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

GC-MS is not the best method. In recent years, LC-MS [19, 21] and shotgun lipidomics [22] have been widely used for the identification and quantification of CEs because of its high resolution, sensitivity, and accuracy. Because of the lack of commercial CE standards and low ionization efficiency, most of the identification and quantification methods have covered only high-abundant CEs (0.01-1μmol/ml) in biological samples. Analytical methods with higher sensitivity and coverage for the global profiling and quantification of CEs are required to evaluate their impact on disease development as well as to discover new biomarkers. In this study, we established a mathematical model-assisted UHPLC-MS method for the untargeted identification to targeted quantification of CEs in plasma, very low density lipoproteins (VLDL), LDL, and HDL samples of humans, rats, normal and hyperlipidemic golden hamsters. To profile unknown CE compounds, we first optimized the ultraperformance liquid chromatography-mass spectrometer (UHPLC-MS) and CE extraction conditions. And an in-house database of theoretically possible CEs was created including the computationally generated precursor ion m/z mass, and a fatty acyl chain with carbon number and double-bond numbers. Then, we used UHPLC-quadrupole-time-of-flight mass spectrometry (UHPLC-QTOF-MS) to detect CE candidates in bio-samples and identify them by comparing with the in-house database. Mathematical models based on chromatographic retention behavior were established and validated by commercial and synthetic standards. A novel strategy was proposed to globally profile and identify CEs by using mathematical model-assisted UHPLC-MS/MS, especially unknown CEs. Finally, the quantification of CEs was performed on dynamic multiple reaction monitoring (dMRM) mode, and used to investigate the distribution of CEs in different species as well as the discovery of new potential biomarkers for hyperlipidemia.

Experimental procedures Materials. Saturated, unsaturated CE standards, and 2 internal standards (Table S1) were purchased from SigmaAldrich (St. Louis, MO, USA), Nu-Chek Prep (Elysian, MN, USA) and Avanti Polar Lipids (Alabaster, AL, USA), respectively. Cholesterol, docosanoic acid (C22:0), tetracosanoic acid (C24:0), heptadecanoic acid (C17:0), 9Znonadecenoic acid (C19:1), 13Z,16Z-docosadienoic acid (C22:2), and 13,16,19-docosatrienoic acid (C22:3) were purchased from J&K Scientific Ltd (Beijing, China), to synthesize CE (22:0), CE (24:0), CE (17:0), CE (19:1), CE (22:2), and CE (22:3) by our laboratory as previously reported [23]. HRMS, 1H-NMR, and 13C-NMR spectra are shown in the Chemical Synthesis of Supplementary Information. Stock solutions of all CEs were prepared in MeOH-CHCl3 (1:1, v/v) at a concentration of 1 mg/mL for each and stored at −80 °C. CE (15:0-d6) and CE (16:0-d6) were dissolved in methanol at a final concentration of 200 ng/mL to obtain the IS solution. The nomenclature of CEs is based on HMDB according to carbon number and degrees of unsaturation of FA, for example CE (18:2) meaning the carbon number of FA is 18 and degrees of unsaturation is 1. UHPLC-MS/MS analysis. The identification process was performed on an Agilent 1290 series UHPLC system coupled to an Agilent 6550 QTOF MS (Agilent Technologies, Santa Clara, CA, USA) with a dual Agilent jet stream electrospray ionization source. The quantification of CEs was performed on an Agilent 1290 series UHPLC system coupled to an Agilent 6470 triple-quadrupole mass spectrometer with the same

Page 2 of 9

ionization device in dynamic MRM mode. Data acquisition and processing were performed by Agilent MassHunter workstation software (version B.07.01). The chromatographic and mass spectrometry conditions of CEs identification and quantification were the same. Data processing and statistical analysis. Agilent MassHunter Qualitative Analysis and Quantitative Analysis (version B.07.01) were used for data processing. Profiling and identification of CEs was simultaneously based on three conditions: (I) high resolution MS (HRMS) data extracted by molecular feature extraction (MFE); (II) MS/MS data containing characteristic product ion (m/z 369.3505) with the highest abundance; and (III) matching the mathematical models built by the chromatographic retention factor (k), the carbon number and degrees of unsaturation of the fatty acyl chain in CEs. The relative retention factor (k) of CE compounds were calculated. With a retention time of CE (18:2) (tm) as a reference, the k of CE (tR) was calculated by k = tR/tm, where tR is the retention time of the CE compound. For the quantification of CEs, the peak area ratios between all CEs and IS were plotted against the real concentrations to construct calibration curves for each CE by the least-squares method with a 1/x weighting factor. If the carbon number of the fatty acyl chain was less than 16, CE (15:0-d6) was selected as IS for the calculation; otherwise, CE (15:0-d6) was selected. CEs without commercial standards were quantified by calibration curve of the standard with the closest structure.

Multivariate statistical analysis was performed by SIMCA-P (version 13.0, Umetrics AB, Umeå, Sweden), in which the PCA model was used to visualize the distance and relevance between the control (Con) and hyperlipidemic model (Mod) groups. The discovery of potential lipid biomarkers required a significant difference where P < 0.05 (bilateral t-test by IBM SPSS Statistics, Version 21, Armonk, USA) and VIP > 1, Jack-knife > 0, and absolute value of Pcorr > 0.58 in the OPLS-DA model. More detailed experimental procedures can be found in the Supplementary information.

Results and discussion To develop a sensitive and accurate method for the global profile and identification of CEs, 17 CE standards were used for optimization of the UHPLC-MS/MS conditions and four extraction systems. First, the MRM parameters, including transition, collision energy, and fragmentor voltage, were optimized using these 17 CE standards and 2 internal standards (IS). The optimized results are listed in Table S1. Then, we comprehensively optimize the mass spectrometry parameters (Fig. S1), including gas temperature and flow, sheath gas temperature and flow, nebulizer pressure, and capillary voltage. Interestingly, the gas temperature is closely related to the abundance response of CEs, because an increase in temperature can degrade the ester bonds of CEs, especially polyunsaturated CEs such as CE (18:2). Next, the liquid chromatographic conditions were optimized. We compared different chromatographic columns, including C8, C18, and C30 reversed-phase columns, as well as solvent systems, including methanol, 2-isopropanol (IPA), and acetonitrile (ACN). As shown in Fig. S2, the UHPLC C18 column (Waters) and IPAACN (5:2, v/v) were chosen based on the balance between the separation resolution and retention time. Four extraction solvents, including Folch, Bligh and Dyer, MeOH-MTBE, and CHCl3-MeOH (10:1, v/v) systems were compared. Fig. S3 shows that the MeOH-MTBE system was the most effective for the extraction of CEs.


2

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

Analytical Chemistry Global profiling and identification of CEs Global profiling of CEs by QTOF-MS. Due to lack of CE standards, high-resolution mass spectrometry providing both HRMS and HRMS/MS data is the best choice for the global profiling and identification of novel CEs in complex biosamples. The fragmentation pattern of CE standards can provide a vital insight into the discovery and identification of unknown CEs. Thus, we summarized the characteristic fragmentation of CEs by QTOF analysis. Fig. 1A shows that

CEs produce the most dominant product ion [Chol+H-H2O] + at m/z 369.3505 formed by losing the fatty acyl chain fragment, as previously reported [19].

Fig. 1. HRMS/MS analysis. (A) Fragmentation of representative CEs by UHPLC-QTOF MS via collision-induced dissociation (CID). (B) Extraction ion chromatography (EIC) of representative CE standards (n = 0, 1, 2). (C) EIC of representative CE compounds (n = 0, 1, 2) identified in the plasma of golden hamsters. Both blue and black peaks represent CEs reported in the HMDB, while blue peaks represent CEs identified by CE standards, and red peaks represent newly identified CEs.


3


Since only the fatty acyl chain of CEs is variable, the m/z at [M+NH4] + of potential CEs can be theoretically predicted if the molecular formula of the fatty acyl chain is known. An inhouse database of theoretically possible CEs was created, including the computationally generated precursor ion m/z mass (Supplementary Excel 1), a fatty acyl chain with carbon number ranging from 2 to 30, and double-bond numbers ranging from 0 to 10. These were then used to probe CEs based on collected high-resolution quasi-molecular ion data by UHPLC-QTOF-MS. Only peaks with a high-resolution mass error less than 15 ppm compared with the theoretical precursor ion m/z value listed in Supplementary Excel 1, as well as with the unique product ion at m/z 369.3505, were accepted as CE candidates. The extraction ion chromatography (EIC) of representative CE standards (n = 0, 1, 2) and CE candidates (n = 0, 1, 2) detected in the plasma of golden hamsters are shown in Fig. 1B-1C. And Fig. S4 shows the HRMS/MS spectrums. Mathematical model-assisted identification of CEs. Based on a careful analysis of characteristics of the CE candidates, the carbon numbers of the fatty acyl chain with the same degree of unsaturation showed a linear relation with retention behavior (lgk) on the C18 RP column. As shown in Fig. 2A, all the regression coefficient (R2) values were greater than 0.99. These results suggest that CEs (n = 0, 1, 2, 3) can be determined by their RT. Using these four linear regression equations, 34 unknown CEs in the plasma of golden hamsters were identified by their lgk values (Fig. 2A). The in-depth analysis in Fig. 2A demonstrates that the degrees of unsaturation of the fatty acyl chain with the same carbon number is closed related to the lgk. Therefore, we constructed prediction curves based on the degrees of unsaturation (Fig. 2B). Similarly, the degrees of unsaturation also had a linear correlation with the lgk. Thus, 8 new polyunsaturated CEs in the plasma of golden hamsters were identified (Fig. 2B). Using the carbon number and degrees of unsaturation rules, 42 new CE candidates were identified. However, it is difficult to identify polyunsaturated CE candidates without at least 3 CE standards with the same carbon number. Therefore, we applied the retention behavior and mass data of the above-confirmed CEs to construct a three dimensional (3-D) mathematical model to profile unknown polyunsaturated CEs. As shown in Fig. 2C, the 3-D scatter plots were constructed using three factors: lgk, carbon number and degrees of unsaturation of fatty acyl chain from the confirmed CEs. These three factors were fitted with the plane regression equation: y = - 0.0661 x1 + 0.0355 x2 + 0.4928 (R2 = 0.9927), where y represented lgk, x1 represented the degrees of unsaturation (n), and x2 represented the carbon number of the fatty acyl chain. We found a close correlation (Fig. 2C, pink plane) between the three factors. To verify the reliability of the 3-D model, we synthesized CE (19:1). A comparison of the measured and calculated lgk value indicated the calculated accuracy value was 103.1%, demonstrating the 3-D model is suitable for the profiling of unknown CEs. Using this 3-D model, 8 novel CEs in the plasma of golden hamsters were profiled. Detailed information of 74 CEs detected (50 newly reported CEs) in the plasma of golden hamsters based on mathematical models are shown in Table S3. The putatively accuracy of the 74 CEs were calculated by the ratio between the measured lgk and the calculated lgk, and the accuracies of the 74 CEs were within 96.8% -104.4%. Deep identification of CEs and validation of mathematical models by pure standards. The detected CEs were deeply identified by using commercially available

standards. Seventeen CEs were confirmed based on the retention time and fragmentation ion. Besides, we synthesized 6 CEs, including CE (19:1), for further structure identification. Extraction ion chromatography (EIC) of representative CE standards (n = 0, 1, 2) are shown in Fig. 1B. Furthermore, 23

Fig. 2. Mathematical model-assisted global profiling of CEs. (A) Regression curve of the measured lgk (k, retention factor from UHPLC-QTOF-MS analysis) versus the carbon numbers of fatty acyl chain (FA) in CEs (n = 0, 1, 2, 3). (B) Regression curve of the measured lgk versus the degrees of unsaturation of FA in CEs (carbon number =18, 20, 22). (C) 3-D model fitting by plane surface fit based on lgk, carbon number and degrees of unsaturation of FA. Blue dots represent CEs identified by commercial and synthetic CE standards (A/B) or the identified4 ACS Paragon Plus Environment CEs in model A and model B (C), and red dots represent the newly identified CEs.

Page 4 of 9

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


Fig. 3. Global profiling of CEs in biological samples. (A) Venn diagrams of CEs detected in plasma, VLDL, LDL, and HDL of humans, rats, and golden hamsters. (B) Distribution of different CE classes with various lengths and saturations of fatty acyl chain in the plasma, VLDL, LDL, and HDL samples of humans, rats, and golden hamsters. (C) Distribution of different CE classes with various lengths and saturations of fatty acyl chain in the plasma, VLDL, LDL, and HDL samples of normal and hyperlipidemic golden hamsters. (D) PCA results of CE concentration data of plasma, VLDL, LDL and HDL samples to assess the degree of dispersion between the Con (Black) and Mod (Red) groups. The two panels represent the top two principal components for the Con and Mod groups.

CE standards were used to validate the efficiency of the mathematical models (Fig. 2A-2B, blue dots). The mathematical models had an excellent prediction performance for CE profiling and identification.

Quantification of CEs in hyperlipidemic hamsters Method validation for quantification. All CEs detected in the biological samples of humans, rats, and golden hamsters by UHPLC-QTOF-MS were transferred to UHPLC-QQQ-MS for

targeted quantification in dynamic MRM mode. Because the two LC-MS/MS systems from Agilent have the same type of UHPLC system and ion source, the MRM parameters (precursor ions, product ions, and CEs) of all identified CEs can be directly determined from CID MS/MS data acquired in QTOF-MS profiling. The detailed MRM parameters and the retention times of the 81 CEs are listed in Table S3, and the MRM spectrum showed in the Fig. S5. The novel CEs quantification method was carefully validated


5


Page 6 of 9

Fig. 4. Box plots of CE potential biomarkers and lipids profile. (A) Box plots of top 10 shared CE potential biomarkers with the largest variable importance in projection values (VIP) in OPLS-DA analysis in plasma samples of the Con and Mod groups. (B) Box plots of lipid profile of the Con and Mod groups. Mod group versus Con group: * P < 0.05, ** P < 0.01, *** P < 0.001. Black CEs were reported in HMDB, and red CEs were newly reported CEs as potential biomarkers for hyperlipidemia.

in terms of sensitivity, linearity, accuracy, precision, matrix effect, recovery, and post-preparative stability using the 17 CE standards with different concentrations (Table S1-S2). The limits of detection (LOD) were 1 ng/mL for all CEs with a signal-to-noise ratio higher than 3. All CEs showed significant calibration linearity, with good correlation coefficients (R2 > 0.99) and a wide dynamic range from 10 to 2000 ng/mL (Table S1). Calibration standards were prepared as described in the Experimental procedures of Supplementary Information. The intra-accuracy and precision of the method were investigated with different quality control (QC) samples of different concentrations by performing six replicates. The interday accuracy and precision was performed on three separate days. The accuracies of low QC, medium QC, and high QC samples were all between 89.9% and 111%, and the RSDs of precisions were less than 12.4% for all CEs (Table S2). Ten mg/mL BSA solution was used as a blank matrix to evaluate the matrix effect and recovery. The matrix effect (ME) value was calculated as ME (%) = B/A × 100, where A is the compound peak area of 100 μl pure standard sample without blank matrix and B is the compound peak area of 100 μl blank matrix spike with 100 μl standard sample after extraction. The recovery value was calculated as R (%) = C/B × 100, where C is the compound peak area of 100 μl blank matrix spike with 100 μl standard sample before extraction. The matrix effect low QC, medium QC, and high QC samples were all between 90.2% and 107%, and the extraction recovery of all QC samples were between 89.4% and 111%. In addition, we also evaluated the post-preparative stability at 4 ℃ for 24 h and 48 h. As shown in

Table S2, the 24 h stability was between 90.4% and 110%, and the 48 h stability was between 91.4% and 110%. The criteria for acceptability of the data included accuracy, matrix effect, recovery, and stability within 80%–120% and precision with an RSD of less than 20%. Good linearity, accuracy, precision, matrix effect, recovery, and post-preparative stability indicates the reliability of CE quantification. This novel method is suitable for the quantification of CEs in biological samples. Golden hamster as the more typical hyperlipidemic model. High-fat diet (HFD) induced hyperlipidemic rats and golden hamsters are classical animal models of hyperlipidemia. TC, LDL-C, and HDL-C, referred to as the total amount of CEs and Chol, are the basic diagnostic indices of hyperlipidemia. To choose a more appropriate hyperlipidemic animal model, we comprehensively compared the components and distribution of CEs in plasma, VLDL, LDL, and HDL of humans, rats, and golden hamsters. Golden hamsters and humans shared more CE compounds in plasma, VLDL, and LDL samples compared with that between rats and humans (Fig. 3A). We further compared the distribution of CEs with different lengths and degrees of unsaturation of FA. CEs are classified by the length and degrees of unsaturation of the only variable fatty acyl chain. Medium-chain (M), long-chain (L), and very-long-chain (VL) CEs contain 6 –12 carbons, 13 – 21 carbons, and more than 22 carbons in their fatty acyl chain, respectively [24, 25]. According to the number of C = C, the degrees of unsaturation of CEs is classified as saturated (n = 0), monounsaturated (n = 1), and


6

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


Table 1. Identified CEs as potential biomarkers link to metabolic diseases Compounds

Link to metabolic diseases

CE (14:0)

Coronary artery disease

CE (16:0)

Cardiovascular disease [14]

[26]

, Hyperlipidemia

, Hyperlipidemia

[14]

, Coronary heart disease

[31]

CE (18:0)

Hyperlipidemia

CE (18:1)

Hypercholesterolemia

CE (18:2)

Metabolic syndrome

[27]

, Diabetes

[28]

[16]

[29]

, Metabolic syndrome

[29]

, Hypercholesterolemia

, Chronic Kidney Disease

[30]


[30]

[26]

[16]

CE (20:0)

Alzheimer Disease

CE (20:3)

Coronary heart disease

CE (20:4)


[26]


[30]


[26]


[30]

[30]


, Non-alcoholic fatty liver disease


, Hypercholesterolemia


[16]


CE (18:3)

CE (22:5)

[14]


[32]

[30]

[32]

[32]


[32]

[34] [31]


polyunsaturated (n ≥ 2). Fig. 3B shows that the distribution of different CEs in plasma and HDL of golden hamsters was similar to that in humans, but those in rats were different with humans. Therefore, with regards to the component and distribution of CEs, golden hamsters are more suitable as the hyperlipidemic animal model than rats. Abnormally elevated CEs in hyperlipidemic hamsters. Here, golden hamsters induced with HFD were used to evaluate abnormal levels of CEs in the hyperlipidemic state. After 14 weeks of feeding with normal (Con group) or HFD diet (Mod group), the levels of TG, TC, LDL-C, and HDL-C in the HFD golden hamsters were significantly increased versus the Con group (Fig. 4B, P < 0.05). In addition, liver histopathological analysis of HFD hamsters showed the significant formation of steatosis and lipid droplet accumulation versus the Con group (Fig. S6). Those results indicated that hyperlipidemia was induced in the HFD hamsters. To investigate the variation of CEs in the hyperlipidemic state, fasting plasma, VLDL, LDL, and HDL samples were obtained and analyzed by the newly established UHPLCMS/MS method. A total of 74, 74, 76, and 73 CEs were identified and quantified in the plasma, VLDL, LDL, and HDL of golden hamsters, respectively. After quantification analysis, Fig. 3C shows that the total amount of all CEs in plasma, VLDL, LDL, and HDL were significantly elevated in hyperlipidemic hamsters compared with the Con group (P < 0.001). Compared to the TC, LDL-C, and HDL-C results by biochemical analysis (P < 0.05), the total amount of CEs in the plasma, LDL, and HDL showed more significant change. In terms of different CE classes, the amount of saturated LCE, unsaturated LCE and VLCE in the plasma and VLDL of the Mod group were significantly elevated. In LDL, the amount of saturated LCE and polyunsaturated VLCE of the Mod group was markedly elevated, and that of LCE in HDL showed a similar trend. Principal component analysis (PCA) revealed that CEs from the plasma, VLDL, LDL, and HDL were clustered separately in the Con and Mod groups (Fig. 3D1-3D4). For in-depth analysis of CE compounds, OPLS-DA models were used for potential biomarker discovery for hyperlipidemia. A total of 57, 52, 42, and 41 CEs were discovered as potential biomarkers in the plasma, VLDL, LDL, and HDL, respectively, as they could discriminate between the Con and hyperlipidemic groups. Heatmap analysis of these CEs is shown in Fig. S7. Among these potential biomarkers, 28, 19, and 14 CEs (21 newly reported CEs of shared 35 CEs) in the VLDL, LDL, and HDL

[30]

, Non-alcoholic fatty liver disease , Non-alcoholic fatty liver disease

[32]

[32]

showed a similar trend to that in plasma of the hyperlipidemic group versus the Con group (Fig. S7, red box), respectively. To further analyze the 35 shared potential biomarkers, box plots of the top 10 CEs with the largest VIP values in plasma were analyzed (Fig. 4A). Compared with the TC, LDL-C, and HDL-C (Fig. 4B), CE (18:0), CE (25:1), CE (14:0), CE (20:0), CE (22:5), and CE (24:1) showed more significant changes (P < 0.001). In addition, the sensitivity of the LC-MS/MS method (1 ng/ml) was higher than that of the biological analysis (0.1 μmol/ml). Interestingly, these potential CE biomarkers were also detected in human samples. Our results indicated that CEs might be effective biomarkers for the diagnosis and therapeutic effect monitoring of hyperlipidemia and hypercholesterolemia. Among the 35 shared potential biomarkers, 14 CEs were reported in HMDB, and 21 CEs were newly reported. These compounds, including at least 11 CEs, were reported to be closely related to diseases (Table 1). CE (14:0), CE (16:0), and CE (18:0) are biomarkers of hyperlipidemia. CE (18:1), CE (18:2), CE (18:3), CE (20:3), CE (22:5), and CE (22:6) are associated with non-alcoholic fatty liver disease, and other CEs are related to diabetes and Alzheimer’s disease, as well as heart diseases including cardiovascular disease and coronary artery disease. CEs, as a “reservoir”, are responsible for cholesterol storage in cells and maintaining the balance of cholesterol supply and metabolism. In our study, we found that the levels of CEs in hyperlipidemic hamsters induced with HFD were significantly elevated. This suggests abnormally elevated or even overloaded CEs in vivo result in metabolic disorders. Traditionally, precursor ion scan (PIS) of [Chol+H-H2O]+ at m/z 369 is used for CE profiling [19]. Agilent 6470 QQQ LC/MS system has been tried to probe CEs in plasma of human and golden hamster by PIS under the same UHPLC condition and MS parameters, and a total of 127 CEs were detected (see Supplementary Excel 3 and PIS MS spectrum). Forty-eight endogenous CEs with low intensity (Area < 1500) in plasma were detected by QQQ because of its higher sensitivity than QTOF (see Supplementary Excel 3). The results demonstrated that the precursor ion scan based on the in-house database can be used to either high-resolution mass spectrometer or tandem mass spectrometer with high sensitivity. And all different types of mass spectrometers could be tried to probe unknown CEs as more as possible. In addition, we tried to create a data list of oxidized CEs based on online database of hydroxyl and hydroperoxy fatty acids on the Lipidmaps database (https://www.lipidmaps.org/),


7


showed in Supplementary Excel 2. Based on the in-house data list, 24 oxidized CEs were tentatively identified due to lack of commercial oxidized CE standards (Supplementary Excel 3). These unknown CEs need to further investigate through synthesizing standards, especially oxidized CEs

Conclusions In summary, we have established a new strategy for the global profiling and identification to targeted quantification of CEs using mathematical model-assisted LC-MS/MS analysis. This new strategy combines chemical synthesis, mathematical models, an animal model, and LC-MS/MS analysis. In this interdisciplinary study, we successfully profiled 81 CEs in plasma samples of humans, rats, and golden hamsters, of which 57 CEs are newly reported. This indicates the high potential for the discovery and identification of unknown CEs. Furthermore, this method was applied for the discovery of potential biomarkers in hyperlipidemic golden hamsters. A total of 57, 52, 42, and 41 CEs were indicated as potential biomarkers in the plasma, VLDL, LDL, and HDL of hyperlipidemic golden hamsters, respectively, and at least 21 shared novel CEs. The in-depth investigation of shared potential biomarkers found that at least 11 CEs were reported to be closely related to metabolic disorders (hyperlipidemia) and heart diseases. Our strategy expands the scope of CE compound analysis in biological samples and can be applied for the discovery of biomarkers for human diseases.

[6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24]

Associated Content

[25]

The Supporting Information is available free of charge on the ACS Publications website. Tables as described in the text and excel. Experimental procedures and Fig. S1−S8 as described in the text (pdf). The PSI MS spectrum was in a single pdf.

[27] [28]

Author Information

[29]

* Corresponding Author Prof. Jin-Lan Zhang ORCID: 0000-0002-6125-7964 Tel.: +86-10-83154880. Fax: +86-10-63017757. E-mail: [email protected].

Author Contributions # These

[26]

[30] [31] [32]

Page 8 of 9

Gleeson.; Richard. J. Deckelbaum.; Gordana Aljinovic. Science. 1996, 272(5266), 1353-1356. Francone. O. L.; Gurakar. A., and Fielding. C. J Biol Chem.1989, 264(12), 7066-7072. Korber. M.; Klein. I.; Daum. G. Biochim Biophys Acta. 2017, 1862(12), 1534. Anderson. R. A.; Byrum. R. S.; Coates. P. M.; Sando. G. N. Proc Natl Acad Sci. 91(7), 2718-2722. Goldstein. J. L.; and Brown. M. S. Cell. 2015, 161(1), 161-172. Harkewicz. R.; Hartvigsen. K.; Almazan. F.; Dennis. E.A.; Witztum. J.L.; Miller. Y.I. J. Biol. Chem. 2008, 283, 1024110251. Wood. W. G.; Li. L.; Muller. W. E.; Eckert. G. P.; J Neurochem. 2014, 129(4), 559-572. Yue. S.; Li. J.; Lee. S.-Y.; Lee. H.J.; Shao. T.; Song. B.; Cheng. L.; Masterson. T.A.; Liu. X.; Ratliff. T.L.; Cheng. J.-X. Cell Metab. 2014, 19, 393-406. Peck. B.; Schulze. A. Cell Metab. 2014, 19, 350-352. Rai S.; Bhatnagar S. OMICS. 2017, 21(3), 132-142. Santos. R. D.; Chacra. A. P.; Morikawa. A. T.; Vinagre. C. C.; Maranhao. R. C. Lipids. 2005, 40(7), 737-743. Padró T.; Cubedo J.; Camino S. J Am Coll Cardiol. 2017, 70 (2), 165-178. Lemieux I.; Lamarche B.; Couillard C. Arch Intern Med (Chic). 2001, 161(22), 2685. Son H H.; Moon J Y.; Seo H S. J Lipid Res. 2014, 55(1), 155162. Yu S.; Dong J.; Zhou W.J Chromatogr B. 2014, 960(6), 222-229. Vahabi F.; Sadeghi S.; Arjmand M. Iran J Basic Med Sci. 2014, 20(7), 835-840. Li. M.; Tong. X.; Lv. P.; Feng. B.; Yang. L.; Liu. H. J Chromatogr A. 2014, 1372, 110-119. Han X. Nat Rev Endocrinol. 2016, 12 (11), 668-679. Corbin. K. D.; Zeisel. S. H. Curr Opin Gastroenterol. 2012, 28(2), 159-165. Touma. E. H.; Charpentier. C. Arch Dis Child. 1992, 67(1), 142145. Beermann. C.; Jelinek. J.; Reinecker. T.; Hauenschild. A.; Boehm. G.; Klör. H.-U. Lipids Health Dis. 2003, 2, 10. Doi: 10.1186/1476-511X-2-10. Tarasov. K. V.; Ekroos. K.; Suoniemi. M.; Kauhanen. D.; Sylvanne. T.; Hurme. R.; Marz. W. J Clin Endocrinol Metab. 2014, 99(1):E45–E52. doi: 10.1210/jc.2013-2559. Stegemann. C.; Pechlaner. R.; Willeit. P.; Langley. S. R.; Mangino. M.; Mayr. U.; Mayr. M. Circulation. 2014, 129(18), 1821-1831. Petersen. KS.; Keogh. JB.; Lister. N.; Weir. JM.; Meikle. PJ.; Clifton. PM. World J Diabetes. 2017, 8(5), 202-212. El-Najjar. N.; Orsó. E.; Wallner. S.; Liebisch. G.; Schmitz. G.; PLoS One. 2014, 10(10), DOI: 10.1371/journal.pone.0140683. Afshinnia. F.; Rajendiran. TM.; Karnovsky. A. Kidney Int Rep. 2016,1(4), 256-268. Wang. L.; Folsom. AR.; Eckfeldt. JH. Nutr Metab Cardiovasc Dis. 2003, 13(5), 256-266. Chan. RB.; Oliveira. TG.; Cortes. EP. J Biol Chem. 2012, 287(4), 2678-2688.

authors have contributed equally to this work.

Notes The authors declare no competing financial interest.

Acknowledgements We gratefully acknowledge financial support from CAMS Innovation Fund for Medical Sciences [Grant number 2016-I2M3-010 and 2018PT35002], and the Drug Innovation Major Project of China (2018ZX09721002-001-004).

References [1] [2] [3] [4] [5]

Ren. J.; Franklin. ET.; Xia. Y. J Am Soc Mass Spectrom. 2017, 28(7), 1432-1441. Horton. J. D.; Goldstein. J. L.; Brown. M. S. Journal of Clinical Investigation. 2002, 109(9), 1125-1131. Hegele. R. A. Nat Rev Genet. 2009, 10(2), 109-121. Suckling. K. E.; and Stange. E. F. J Lipid Res. 1985, 26(6), 647671. Hongyuan Yang.; Martin Bard.; Debora. A. Bruner.; Anne.


8

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


Only for TOC 283x115mm (300 x 300 DPI)


MS ... - ACS Publications

Recommend Documents