Applying lipidomics strategies to study lipid metabolic diseases
Zhang, Wenxuan
DOI:
10.33612/diss.169407826
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2021
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Zhang, W. (2021). Applying lipidomics strategies to study lipid metabolic diseases. University of Groningen. https://doi.org/10.33612/diss.169407826
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
7
Chapter 7
Lipidomics should still be considered as an emerging “-omics” field in view of the ongoing rapid development of technical possibilities and its potential use for future applications in basic and applied research. The current technical “routine” allows us to extract and detect the most common lipid species with a certain identification level concerning their structures. However, new lipidomics workflows are still in demand when information is required on the detailed structure and composition of lipids, their distribution over the different organs, the constituting cell populations and even the different subcellular organelles. This information has been proven to be vital in understanding the pathology of certain diseases, such as chronic kidney disease and myelin disease1–3. Additionally, standardized computational processing pipelines to handle such complex datasets are still under development. Next to the processing of lipidomics data, another major challenge is to extract useful messages from the huge datasets that are needed to answer different types of biological questions. The aim of the studies presented in this thesis is to establish an untargeted lipidomics workflow and to apply this technique in a series of biomedical studies to identify key lipid changes that may represent cause or consequence of metabolic alterations. In addition, we combined the lipidomics workflow that was developed with other omics workflows (transcriptomics, proteomics) to explore the possibilities of combining the information extracted from lipid profiles with the other layers of information and to explain (patho)physiological phenotypes in lipid metabolism. In this chapter, I will discuss the major findings of the thesis and share my thoughts on future perspectives and challenges for lipidomics and its application in (pre)clinical studies.
Pre-analytical steps are crucial for the success of lipidomics experiments
The pre-analytical steps in any lipidomics workflow are of vital importance but are often neglected in the description of the experimental design. There are many concerns that require special attention with respect to sample type and the target lipids. In Chapter 2 we discuss the lipids that are potentially affected
by pre-analytical factors during sample preparation, storage and lipid exraction procedures. However, several issues were not addressed in our reports, because some of them only have been recognized recently. One of these issues is the conversion between lipids during sample handling and storage which can strongly affect the outcome of lipidomic analyses. For instance, lysophospholipid species and monoacylcardiolipins in cell materials can be generated artificially from phospholipids and cardiolipins by enzymatic reactions4,5. This effect can be minimized by avoiding sample handling at room temperature and immediate snap-freezing of samples at −80 °C. Isomerization of lysophospholipids was
7
also observed to occur at room temperature, even when the lipids were storedin methanol6. This transition can lead to erroneous results when particular lysophospholipids are being used as standards for quantification purposes. Thus, lipid standards should be aliquoted properly to avoid unnecessary freeze and thaw cycles. Oxidation of unsaturated lipid species may occur through enzymatic and free radical-mediated reactions, so antioxidants, metal chelating agents and reducing agents are considered essential for long-time storage. These procedural details should be taken into consideration when planning an experiment. Another topic which is not covered in our report is the optional derivatization step after lipid extraction. Derivatization is usually applied for targeted and semi-targeted detection of certain lipid species. This additional step in sample work up is designed to improve ionization efficiency for MS detection or to generate lipid derivatives to aid the identification of isobaric lipids in a complex sample. Although detection of certain lipid species can be partially achieved by LC-MS method optimization, in some cases, lipid derivatization helps to improve the quality of the measurements by better separating isobaric lipids chromatographically. Several published derivatization methods have been summarized in a recently published review7.
In Chapter 3, we specifically evaluated several lipid extraction procedures,
including a single phase extraction and two-phase extraction methods of plasma samples, and showed the pros and cons of these methods. Plasma is the preferred bio-fluid for clinical diagnostics and biomarker development. Lipid signatures in plasma show great potential to understand disease-associated phenotypes at the molecular level. We reported the extraction efficiencies of frequently used extraction solvent systems (Folch method88, Bligh and Dyer method9, MTBE method10 and MMC method11) towards different lipid classes. This evaluation helps the selection of the optimal extraction method according to the aim of the specific study. Recently, new or adapted lipid extraction procedures have been developed for lipidomics workflows, yet, we did not include these in our comparisons. Some are protein precipitation-based one phase extraction methods and others are two phase or three phase extraction methods12–15. These new methods either claimed better extraction efficiency for specific sample types or improved compatibility with other -omics measurements such as metabolomics and proteomics. These new methods should be tested and evaluated to be applied in large collections of clinical samples or in multi-omics pipelines.
Potential predictive role of plasma lipid profiles for different genetic causes of extreme LDL-c levels in humans
Abnormally elevated level of plasma LDL-c is a well-established risk factor for atherosclerotic cardiovascular disease in humans. Genetic causes and unhealthy lifestyle are the main underlying reasons for alterations of LDL-c levels16. Among subjects with extreme levels of LDL-c, genetic defects are considered as the major determinants. Although presenting with a similar LDL-c level, the genetic causes underlying abnormal LDL-c may be diverse. Common gene variants affecting LDL-c levels are mutations in apolipoprotein B (APOB), apolipoprotein E (APOE), low density lipoprotein receptor (LDLR), proprotein convertase subtilisin/kexin Type 9 (PCSK9), microsomal triglyceride transfer protein (MTTP) and others17. New or rare potential gene variants, which also result in an abnormal LDL-c level, are continuously being reported18,19. In Chapter 4, we measured plasma lipid profiles of two subsets of participants of the LifeLines population cohort with either extremely high or extremely low LDL-c levels. Within each group with either high or low LDL-c, we observed very different lipid profiles although the subjects shared similar LDL-c levels. Interestingly, in the extremely low LDL-c subjects, the plasma lipid profiles tended to be similar when they shared the same genetic defect in the APOB gene. However, the lipid profiles did not separate according to the genetic backgrounds in the extremely high LDL-c population. A major possible reason is the complexity of the many confounding factors involved. For instance, the use of statins can affect the plasma lipid profiles. Other than genetic background and medication, factors that contribute to high LDL-c, such as smoking and high age, were present in the relatively small population studied (n=23). This brought up the necessity of classifying the subjects prior to analysis to properly minimize the interference of confounding factors. So a larger and “better-defined” population will be required for more accurately testing the efficacy and clinical applicability of plasma lipid profiling to evaluate the mechanism underlying elevated LDL-c levels.
Lipidomic analysis revealed the adaptative strategies of mice with a single genetic defect in fatty acid metabolism
Inborn errors of metabolism (IEMs) are groups of inherited disorders with mutations in genes encoding key enzymes of metabolic pathways. One of the prominent harmful consequences of such mutations is the accumulation of intermediate metabolites when a certain enzyme activity is reduced. Thus, targeted metabolic assays have been developed to detect these circulating metabolites and are widely used for standard diagnostic procedures of IEMs in specialized laboratories around the globe. However, there is still a broad interest
7
in seeking for novel metabolic markers to reduce the rate of false diagnoses orto help predicting the severity of the symptoms over time. Moreover, insight in potential adaptative metabolic strategies in IEM patients may help to explain the well-known diversity in phenotypes: this would be highly instructive for personalized medicine purposes. In this case, a comprehensive untargeted metabolomics approach could provide a broader vision on the consequences of a certain mutation on prevailing concentrations of a series of metabolites. Lipidomics is particularly attractive when studying IEMs related to lipid metabolism. In Chapter 5, we systematically evaluated the lipidomes of different
tissues of MCAD KO mice after fasting and exposure to cold. Interestingly, we found that Triglyceride appeared to serve as a sink preventing the accumulation of toxic medium-chain fatty acids and their secondary metabolic products. This mechanism was found to exist in liver and brown adipose tissue under the applied stress conditions. Some of the medium chain fatty acyl containing TGs were also detected in blood of MCAD KO mice in very high concentrations compared to WT mice. Although the changes of TGs were observed in mice under cold exposure and fasting conditions, similar changes were also seen in BAT of MCAD KO mice under fasting conditions only (unpublished data). This indicates that the “TG sink” may represent a common strategy adopted by mice with a genetic defect in the MCAD enzyme. It will be particularly interesting to test for the presence of these TG species in MCAD patients. It was reported previously that very long chain lysophospholipids were enriched in patients with peroxisomal disorders20. These lipid signatures were also closely correlated with the severity of the peroxisomal disorder and should be evaluated further for diagnostic purposes. Therefore, the medium chain TGs or other lipid signatures in plasma of MCAD patients are worth to be evaluated for their value in diagnostics, as they might have the potential to distinguish patients with mild symptoms from severe cases. Of course, this hypothesis needs to be tested in a relatively large population of patients.
An emerging multi-omics approach to study (dys)regulation of lipid metabolism
The development of novel lipidomics methodologies now offers the chance to measure diverse lipid species, but also offers the possibility to assess the cross-talk with other omics data. This is particularly beneficial in the evaluation of complex metabolic diseases that affect lipid metabolism, such as non-alcoholic fatty liver disease (NAFLD), since the identified lipid species may fill the information gap between altered expression of genes/proteins and clinical phenotypes related to disturbed lipid metabolism. The advantage of a multi-omics strategy lies in the comprehensive evaluation of potential mechanisms
underlying changes in different aspects of lipid metabolism. Evidence collected from multiple layers of information will also improve the reliability of the observations. For a complex disease phenotype that is driven by multiple factors, it is very challenging to decipher the most affected signaling pathways or the underlying molecular mechanisms. Hypothesis-based classical biochemical measurements provide targeted observations, yet, they are biased by prior knowledge and (personal) insights. Single untargeted omics approaches that measure a broader range of molecular features are useful but sometimes prove insufficient to set focus due to an overwhelming amount of changes sorted from data without biological explanations. Cross-talk between different -omics layers will likely improve the efficiency of information sorting when subsets of changes captured at different levels are supporting each other . In an effort of integrating lipidomics with other –omics techniques, in Chapter 6 we
explored the regulatory role of the transcription factor LRH-1 on hepatic lipid metabolism in an Lrh-1 knockdown mouse model challenged with a Western-type diet known to induce NAFLD in mice. We combined the information from gene transcripts, proteins and lipids through omics integration models and pathway networking and, thereby, were able to unravel new links between the regulatory networks of LRH-1 and hepatic lipid metabolism. The validity of this approach was proven by the identification of well-known targets of LRH-1 such as Scarb1 and Fads2 as determined in other studies with LRH-1-deficient mouse models21,22. This validation step emphasized the general reliability of the multi-omics integrative strategy and its value for hypothesis generation. The newly identified regulatory axis of LRH-1 and sphingolipid biosynthesis reported in our study designated a distinct role of LRH-1 in controlling ceramide production via ORM-like (ORMDL) proteins23.
A multi-disciplinary study as described in Chapter 6 offers an individual
researcher the opportunity to access large amounts of information and also calls for close collaboration between researchers and extensive knowledge exchange. Untargeted omics techniques and the corresponding “big-data” are still not very familiar for clinical scientists and for a large proportion of biomedical researchers. On the other side, researchers active in the various -omics fields are generally struggling with choices concerning optimal ways of organizing the data and, particularly, with biological interpretations of their findings. Therefore, multi-disciplinary collaborations are recommended to advance this kind of studies to a next level and to help in translating findings to mechanistic insights and, eventually, to clinical applications.
7
Current applications of clinical lipidomics
In this thesis, I described the considerable potential of applying lipidomics for disease phenotyping, identifying metabolic adaptations and multi-omics integrations. One of the ultimate goals is to apply lipidomics in clinical diagnostics including the development of lipid biomarkers and their use in clinical phenotyping. Determination of disease stage, severity and subtype is very important for precise characterization of the individual patient and for design of a personalized treatment strategy. Traditional lipid measurements such as plasma TG and LDL-c levels have been used for decades as risk factors or indications of cardiovascular disease24, yet, the value of single lipid species or disease-related lipid profiles have not been fully appreciated. Current clinical application of lipidomics has a strong focus on cohort comparison. Cohorts of patients with well-described clinical phenotypes are compared with apparently healthy volunteers. Next to this type of studies, associations have been explored between differences in lipid profiles and severity or heterogeneity in cohorts of patients with complex diseases, like the study presented in chapter 4. Xiangdong Wang and coworkers recently merged clinical phenomics with lipidomics in several multi-omics studies in patients with various lung diseases and lung cancer after clinical diagnosis25–27. They analyzed the individual lipid species in plasma samples of a cohort of patients with various forms of critical pulmonary illnesses. Diagnostic accuracy of the identified lipid species was evaluated by receiver operating characteristic (ROC) analysis of the individual species. The area under the curve (AUC) was used to evaluate the diagnostic accuracy and was found to vary between 0.64 for PG40:5 and 0.86 (significant at P<0.05) for PS34:2, compared to a maximal variability of the AUC in ROC analysis of 0.50 – 1.00. They did not test whether diagnostic accuracy improved by taking combinations of lipid species.
Gene-lipid linkages in plasma are now also being explored28,29. Genome-wide association studies (GWAS) were performed in large populations of humans or mice from diversity outbred strains or multiparent populations derived from highly diverse founder strains whose plasma lipidome was also measured. The resulting databases based on Quantitative trait locus (QTL) mapping are becoming publicly available, some of which can be approached from either side, i.e., from the gene-centered or the lipid-centered view30. Plasma lipid-species-associated loci were identified and were also proven to associate with metabolic traits such as body mass index (BMI) and cardiovascular disease (CVD) risk. These data provide new understanding of genetic control
of lipid molecules together with potential prediction ability of those single lipid species on complex metabolic phenotypes29.
Screening and assessing lipid signatures in the studies presented in this thesis as well as published work show the potential of lipidomics to improve the risk prediction of complex diseases. However, translating lipidomics data to clinical lipidomics still faces multiple challenges at the analytical and the post-analytical side.
Challenges and opportunities of Clinical lipidomics
Challenges in analytical aspects
Integrating lipidomics into clinical testing and analysis requires stable and standardized operation procedures to ensure the reliability and the reproducibility of results. As a new branch of omics science, various sample preparation protocols and analytical methods have been developed across different laboratories. Lacking consistency of the procedures sometimes results in poor transferability or comparability of the results between different laboratories. Studies showed that even using the same set of samples and the same extraction and analytical methods, different LC-MS instrumentation systems could lead to slightly different outputs31. All of the above mentioned concerns need to be addressed before applying lipidomics in the clinical work-up of a patient with an undiagnosed disease. What should also be emphasized is the issue to assign the lipid structure in detail by improving both analytical method and data preprocessing pipelines for lipid identification. Current analytical and data processing workflows allow rapid analysis to the level of the lipid species or the constituent fatty acids (Figure 1, orange box). However,
there is still a technical barrier to reach more precise annotation levels such as the position of double bonds in a given fatty acid and their cis/trans geometry in large scale or routine lipidomic studies. However, this information might be vital for biomarker discovery and for establishing the physiological role of the lipids. As a consequence, several recently published works, especially designed to address those issues in more detail32–34, were restricted due to the lack of efficient data processing solutions. Figure 2 summarizes the major
analytical challenges for the current lipidomics workflow. I believe that the development of the technique and the standardization of the protocol will solve the reproducibility and specificity issues and forward a more broad application of clinical lipidomics.
7
Figure 1. Current lipid identification hierarchy according to the resolution of lipid structures.
The top three levels indicate the lipid class and subclass information. The identification levels within orange boxes are the most common identification levels used in current large scale lipidomics analysis. Lipids identified with sub species, isomer and geometric isomer levels usually require additional sample preparation and/or analytical steps. The figure is modified from Foster JM et al with permission. “LipidHome: A Database of Theoretical Lipids Optimized for High Throughput Mass Spectrometry Lipidomics”. PLoS One. 2013;8(5):e61951. doi:10.1371/ journal.pone.006195135
Challenges of post-analytical processing of multi-omics data
In addition to the discovery of lipid biomarkers, the other, more integrative application of clinical lipidomics is to merge lipidomics with other omics data as well as with clinical information to achieve comprehensive clinical phenomes to provide a more precise diagnosis and to understand disease pathology. One of the first steps is to merge acquired lipid information with the known pathways in lipid metabolism. However, this step requires improvements in bio-informatics of metabolic pathways and their regulation next to the above mentioned improvements in lipid identification. For many enzymes involved in lipid metabolism, substrate specificity and mechanism of action are only partially understood. The lipid substrates participating in the various enzymatic reactions are, in general, poorly annotated. Correspondingly, the construction of enzymatic pathways in lipid metabolism in current public pathway databases is still incomplete. For instance, databases which are commonly used, such as the KEGG and the Wiki pathway databases, only include few lipid molecules
at accurate structural levels, while the rest of lipids are annotated at the level of their subclass. So accurately integrating lipids into the current trans-omics functional mappings still needs development of both lipid identification and the advance of biological understanding.
Figure 2. Current analytical and data-preprocessing challenges. Multiple steps are in need of
standardization in order to be used for clinics. The challenges in bold require more attention. Colors of the terms indicate different stages of a lipidomics workflow. Adopted from Liebisch G et al with permission. “Lipidomics needs more standardization”. Nat Metab. 2019;1(8):745-747. doi:10.1038/s42255-019-0094-z 36
Challenges and Opportunities of Clinical lipidomics in patient diagnosis
Application of lipidomics to diagnose a disease and to define a treatment strategy for the given patient is still a major challenge. Diagnosis was commonly established by comparing the analytical results of body fluids (e.g., plasma, urine) of an individual patient with a large dataset of apparently healthy volunteers. For diagnosis of IEMs, metabolites were considered pathognomonic when they resulted in the unequivocal identification of the affected metabolic enzyme reaction, directly linked to the mutated gene. Metabolite-gene associations are thus established by virtue of the possibility to identify abnormal metabolites based on datasets of humans not affected by an IEM and
7
an extensive and detailed knowledge of metabolic pathways. For lipids, bothare lacking to an appreciable extent. The building of datasets of analysis results of many humans over many years is at this moment an important impediment, because required methodologies are basically lacking. Knowledge of lipid metabolism, particularly that of complex lipids, is still at a rather cursory level compared to classical metabolic pathways of other metabolites. As a consequence, automated pipelines to associate lipid species/metabolites with specific genes are lacking. At this moment, a gene-centered approach is usually applied and lipidomics and metabolomics are requested when a mutation has been identified in a gene of unknown function. The project that underpins this thesis and started from the idea to apply multi-omics analysis on a single plasma sample to establish a diagnosis (SysMed) is further away than originally anticipated. But I also see progress in mapping lipid traits on genes and vice
versa, improving detailed knowledge on metabolism of complex lipids and
methods to ‘stitch’ datasets from different population studies together offering us a glimpse of a not so distant future.
Concluding remarks
In the last few years, I have witnessed a rapid development of lipidomics workflows and of a general acceptance of applications of the technique. Researchers are exploring different possibilities of using lipidomics to answer different biological and/or clinical questions. In this thesis, we made our steps in merging lipidomics with different types of lipid-related research. I firmly believe that lipidomics will become more standardized and integrated in different types of studies to solve important puzzles in lipid metabolism and to assist in clinical diagnostics.
References
1. Maganti RJ, Hronowski XL, Dunstan RW, et al. Defining Changes in the Spatial Distribution and Composition of Brain Lipids in the Shiverer and Cuprizone Mouse Models of Myelin Disease. J
Histochem Cytochem. 2019;67(3):203-219. doi:10.1369/0022155418815860
2. Yutuc E, Angelini R, Baumert M, et al. Localization of sterols and oxysterols in mouse brain reveals distinct spatial cholesterol metabolism. Proc Natl Acad Sci U S A. 2020;117(11):5749-5760. doi:10.1073/pnas.1917421117
3. Noh SA, Kim SM, Park SH, et al. Alterations in Lipid Profile of the Aging Kidney Identified by MALDI Imaging Mass Spectrometry. J Proteome Res. 2019;18(7):2803-2812. doi:10.1021/acs. jproteome.9b00108
4. Liebisch G, Schmitz G. Quantification of lysophosphatidylcholine species by high-throughput electrospray ionization tandem mass spectrometry (ESI-MS/MS). Methods Mol Biol. 2009;580:29-37. doi:10.1007/978-1-60761-325-1_3
5. Kim J, Hoppel CL. Comprehensive approach to the quantitative analysis of mitochondrial phospholipids by HPLC-MS. J Chromatogr B Anal Technol Biomed Life Sci. 2013;912:105-114. doi:10.1016/j.jchromb.2012.10.036
6. Okudaira M, Inoue A, Shuto A, et al. Separation and quantification of 2-acyl-1-lysophospholipids and 1-acyl-2-lysophospholipids in biological samples by LC-MS/MS. J Lipid Res. 2014;55(10):2178-2192. doi:10.1194/jlr.D048439
7. Zhao XE, Zhu S, Liu H. Recent progresses of derivatization approaches in the targeted lipidomics analysis by mass spectrometry. J Sep Sci. 2020;43(9-10):1838-1846. doi:10.1002/jssc.201901346 8. Folch J, Lees M, Sloane GH. A SIMPLE METHOD FOR THE ISOLATION AND PURIFICATION OF TOTAL
LIPIDES FROM ANIMAL TISSUES* Downloaded From. Accessed December 30, 2020. http://www.jbc.
org/
9. BLIGH EG, DYER WJ. A rapid method of total lipid extraction and purification. Can J Biochem Physiol. 1959;37(8):911-917. doi:10.1139/o59-099
10. Matyash V, Liebisch G, Kurzchalia T V., Shevchenko A, Schwudke D. Lipid extraction by methyl-terf-butyl ether for high-throughput lipidomics. In: Journal of Lipid Research. Vol 49. ; 2008:1137-1146. doi:10.1194/jlr.D700041-JLR200
11. Pellegrino RM, Di Veroli A, Valeri A, Goracci L, Cruciani G. LC/MS lipid profiling from human serum: A new method for global lipid extraction. Anal Bioanal Chem. 2014;406(30):7937-7948. doi:10.1007/ s00216-014-8255-0
12. Nakayasu ES, Nicora CD, Sims AC, et al. MPLEx: a Robust and Universal Protocol for Single-Sample Integrative Proteomic, Metabolomic, and Lipidomic Analyses. mSystems. 2016;1(3):2020. doi:10.1128/msystems.00043-16
13. Vale G, Martin SA, Mitsche MA, Thompson BM, Eckert KM, McDonald JG. Three-phase liquid extraction: A simple and fast method for lipidomic workflows. J Lipid Res. 2019;60(3):694-706. doi:10.1194/jlr.D090795
14. Satomi Y, Hirayama M, Kobayashi H. One-step lipid extraction for plasma lipidomics analysis by liquid chromatography mass spectrometry. J Chromatogr B Anal Technol Biomed Life Sci. 2017;1063:93-100. doi:10.1016/j.jchromb.2017.08.020
15. Furse S, Koulman A. Lipid extraction from dried blood spots and dried milk spots for untargeted high throughput lipidomics. Mol Omi. 2020;16(6):563-572. doi:10.1039/D0MO00102C
16. Vodnala D, Rubenfire M, Brook RD. Secondary causes of dyslipidemia. Am J Cardiol. 2012;110(6):823-825. doi:10.1016/j.amjcard.2012.04.062
17. Burnett JR, Hooper AJ. Common and rare gene variants affecting plasma LDL cholesterol. Clin
Biochem Rev. 2008;29(1):11-26. http://www.ncbi.nlm.nih.gov/pubmed/18566665%0Ahttp://www.
7
18. Riveros-Mckay F, Oliver-Williams C, Karthikeyan S, et al. The influence of rare variants in circulatingmetabolic biomarkers. PLoS Genet. 2020;16(3):e1008605. doi:10.1371/journal.pgen.1008605 19. Bomba L, Walter K, Soranzo N. The impact of rare and low-frequency genetic variants in common
disease. Genome Biol. 2017;18(1):1-17. doi:10.1186/s13059-017-1212-4
20. Herzog K, Pras-Raves ML, Ferdinandusse S, et al. Plasma lipidomics as a diagnostic tool for peroxisomal disorders. J Inherit Metab Dis. 2018;41(3):489-498. doi:10.1007/s10545-017-0114-7 21. Miranda DA, Krause WC, Cazenave-Gassiot A, et al. LRH-1 regulates hepatic lipid homeostasis and
maintains arachidonoyl phospholipid pools critical for phospholipid diversity. JCI insight. 2018;3(5). doi:10.1172/jci.insight.96151
22. Schoonjans K, Annicotte JS, Huby T, et al. Liver receptor homolog 1 controls the expression of the scavenger receptor class B type I. EMBO Rep. 2002;3(12):1181-1187. doi:10.1093/embo-reports/ kvf238
23. Siow D, Sunkara M, Dunn TM, Morris AJ, Wattenberg B. ORMDL/serine palmitoyltransferase stoichiometry determines effects of ORMDL3 expression on sphingolipid biosynthesis. J Lipid Res. 2015;56(4):898-908. doi:10.1194/jlr.M057539
24. Hokanson JE, Austin MA. Plasma Triglyceride Level Is a Risk Factor for Cardiovascular Disease
Independent of High-Density Lipoprotein Cholesterol Level: A Metaanalysis of Population-Based Prospective Studies. Vol 3.; 1996. doi:10.1177/174182679600300214
25. Lv J, Gao D, Zhang Y, Wu D, Shen L, Wang X. Heterogeneity of lipidomic profiles among lung cancer subtypes of patients. J Cell Mol Med. 2018;22(10):5155-5159. doi:10.1111/jcmm.13782
26. Gao D, Zhang L, Song D, et al. Values of integration between lipidomics and clinical phenomes in patients with acute lung infection, pulmonary embolism, or acute exacerbation of chronic pulmonary diseases: A preliminary study. J Transl Med. 2019;17(1):162. doi:10.1186/s12967-019-1898-z
27. Zhang L, Zhu B, Zeng Y, Shen H, Zhang J, Wang X. Clinical lipidomics in understanding of lung cancer: Opportunity and challenge. Cancer Lett. 2020;470:75-83. doi:10.1016/j.canlet.2019.08.014 28. Jha P, McDevitt MT, Halilbasic E, et al. Genetic Regulation of Plasma Lipid Species and Their Association
with Metabolic Phenotypes. Cell Syst. 2018;6(6):709-721.e6. doi:10.1016/j.cels.2018.05.009 29. Tabassum R, Rämö JT, Ripatti P, et al. Genetic architecture of human plasma lipidome and its link to
cardiovascular disease. Nat Commun. 2019;10(1). doi:10.1038/s41467-019-11954-8
30. Linke V, Overmyer KA, Miller IJ, et al. A large-scale genome–lipid association map guides lipid identification. Nat Metab. 2020;2(10):1149-1162. doi:10.1038/s42255-020-00278-3
31. Cajka T, Smilowitz JT, Fiehn O. Validating Quantitative Untargeted Lipidomics Across Nine Liquid Chromatography-High-Resolution Mass Spectrometry Platforms. Anal Chem. 2017;89(22):12360-12368. doi:10.1021/acs.analchem.7b03404
32. Harris RA, May JC, Stinson CA, Xia Y, McLean JA. Determining Double Bond Position in Lipids Using Online Ozonolysis Coupled to Liquid Chromatography and Ion Mobility-Mass Spectrometry. Anal
Chem. 2018;90(3):1915-1924. doi:10.1021/acs.analchem.7b04007
33. Ma X, Chong L, Tian R, et al. Identification and quantitation of lipid C=C location isomers: A shotgun lipidomics approach enabled by photochemical reaction. Proc Natl Acad Sci U S A. 2016;113(10):2573-2578. doi:10.1073/pnas.1523356113
34. Cao W, Cheng S, Yang J, et al. Large-scale lipid analysis with C=C location and sn-position isomer resolving power. Nat Commun. 2020;11(1):1-11. doi:10.1038/s41467-019-14180-4
35. Foster JM, Moreno P, Fabregat A, et al. LipidHome: A Database of Theoretical Lipids Optimized for High Throughput Mass Spectrometry Lipidomics. Oresic M, ed. PLoS One. 2013;8(5):e61951. doi:10.1371/journal.pone.0061951
36. Liebisch G, Ahrends R, Arita M, et al. Lipidomics needs more standardization. Nat Metab. 2019;1(8):745-747. doi:10.1038/s42255-019-0094-z
