STAR★Methods

Key Resources Table
REAGENT or RESOURCE SOURCE IDENTIFIER
Antibodies
ALIX Abcam Cat#ab76608; RRID: AB_2042595
Peroxidase-Conjugated Goat anti-Rabbit IgG (H+L) Beijing Zhongshan Jinqiao Biotechnology Co., Ltd. Cat#ZB-2301; RRID: AB_2747412
Chemicals, Peptides, and Recombinant Proteins
Chloroform (HPLC grade) Honeywell Cat#049-4
Methanol (HPLC grade) Fisher chemical Cat#A452-4
Acetonitrile (LCMS grade) Fisher chemical Cat#A955-4
Formic acid (98%) J&K Cat#299272
RIPA lysis buffer Meilunbio Cat#MA0151
Protease inhibitor cocktail Sigma-Aldrich Cat#P8340-5ML
Ammonium hydroxide solution Sigma-Aldrich Cat#05002-1L
Ammonium acetate Sigma-Aldrich Cat#73594
PC-14:0/14:0 Avanti Polar Lipids Cat#850345P
d31-PC16:0/18:1 Avanti Polar Lipids Cat#860399C
PE14:0/14:0 Avanti Polar Lipids Cat#850745P
d31-PE-16:0/18:1 Avanti Polar Lipids Cat#860374C
d31-PS-16:0/18:1 Avanti Polar Lipids Cat#860403C
PA-17:0/17:0 Avanti Polar Lipids Cat#830856P
PG-14:0/14:0 Avanti Polar Lipids Cat#840445P
d31-PG-16:0/18:1 Avanti Polar Lipids Cat#860384C
C14:0-BMP Avanti Polar Lipids Cat#857131P
d31-PI-16:0/18:1 Avanti Polar Lipids Cat#860042P
SM-d18:1/12:0 Avanti Polar Lipids Cat#860583P
LPC-17:0 Avanti Polar Lipids Cat#855676P
LPE-17:1 Avanti Polar Lipids Cat#110699
LPI-17:1 Avanti Polar Lipids Cat#850103P
LPA-17:0 Avanti Polar Lipids Cat#857127P
LPS-17:1 Avanti Polar Lipids Cat#858141P
S1P-d17:1 Avanti Polar Lipids Cat#860641P
Cer-d18:1/d7-15:0 Avanti Polar Lipids Cat#860681P
GluCer d18:1/8:0 Avanti Polar Lipids Cat#860540P
GalCer d18:1/8:0 Avanti Polar Lipids Cat#860538P
PI-8:0/8:0 Echelon Cat#P-0008
d3-GM3 d18:1/18:0 Matreya LLC Cat#2052
d3-LacCer d18:1/16:0 Matreya LLC Cat#1534
d5-DAG16:0/16:0 Avanti Polar Lipids Cat#110537
d5-DAG18:1/18:1 Avanti Polar Lipids Cat#110581
TAG(14:0)3-d5 CDN Isotopes Cat#D-6958
TAG(16:0)3-d5 CDN Isotopes Cat#D-5815
TAG(18:0)3-d5 CDN Isotopes Cat#D-5816
d6-cholesterol CDN Isotopes Cat#D-2139
d6-CE18:0 CDN Isotopes Cat#D-5823
L-Phenylalanine-d8 Cambridge Isotope Laboratories Cat# DLM-372-1
L-Tryptophan-d8 Cambridge Isotope Laboratories Cat#DLM-6903-0.25
L-Isoleucine-d10 Cambridge Isotope Laboratories Cat#DLM-141-0.1
L-Asparagine-13C4 Cambridge Isotope Laboratories Cat#CLM-8699-H-0.05
L-Methionine-d3 Cambridge Isotope Laboratories Cat#DLM-431-1
L-Valine-d8 Cambridge Isotope Laboratories Cat#DLM-311-0.5
L-Proline-d7 Cambridge Isotope Laboratories Cat#DLM-487-0.1
L-Alanine-d7 Cambridge Isotope Laboratories Cat#DLM-251-PK
DL-Serine-d3 Cambridge Isotope Laboratories Cat#DLM-1073-1
DL-Glutamic acid-d5 Cambridge Isotope Laboratories Cat#DLM-357-0.25
L-Aspartic acid-d3 Cambridge Isotope Laboratories Cat#DLM-546-0.1
L-Arginine-d7 Cambridge Isotope Laboratories Cat#DLM-541-0.1
L-Glutamine-d5 Cambridge Isotope Laboratories Cat#DLM-1826-0.1
L-Lysine-d9 Cambridge Isotope Laboratories Cat#DLM-570-0.1
L-Histidine-d5 Cambridge Isotope Laboratories Cat#DLM-7855
Taurine-13C 2 Cambridge Isotope Laboratories Cat#CLM-6622-0.25
Betaine-d11 Cambridge Isotope Laboratories Cat#DLM-407-1
Urea-(13C,15N2) Cambridge Isotope Laboratories Cat#CLM-234-0.5
L-lactate-13C3 Sigma-Aldrich Cat#485926-500MG
Trimethylamine N-oxide-d9 Cambridge Isotope Laboratories Cat#DLM-4779-1
Choline-d10 Cambridge Isotope Laboratories Cat#DLM-141-0.1
Malic acid-d3 Cambridge Isotope Laboratories Cat#DLM-9045-0.1
Citric acid-d4 Cambridge Isotope Laboratories Cat#DLM-3487-0.5
Succinic acid-d4 Cambridge Isotope Laboratories Cat#DLM-584-1
Fumaric acid-d4 Cambridge Isotope Laboratories Cat#DLM-7654-1
Hypoxanthine-d3 Cambridge Isotope Laboratories Cat#DLM-2923-0.1
Xanthine-15N2 Cambridge Isotope Laboratories Cat#NLM-1698-0.1
Thymidine (13C10,15N2) Cambridge Isotope Laboratories Cat#CNLM-3902-25
Inosine-15N4 Cambridge Isotope Laboratories Cat#NLM-4264-0.01
Cytidine-13C5 Cambridge Isotope Laboratories Cat#CLM-3679-0.05
Uridine-d2 Cambridge Isotope Laboratories Cat#DLM-7693-0.05
Methylsuccinic acid-d6 Cambridge Isotope Laboratories Cat#DLM-2960-1
Benzoic acid-d5 Cambridge Isotope Laboratories Cat#DLM-122-1
Creatine-d3 Cambridge Isotope Laboratories Cat#DLM-1302-0.25
Creatinine-d3 Cambridge Isotope Laboratories Cat#DLM-3653-0.1
Glutaric acid-d4 Cambridge Isotope Laboratories Cat#DLM-3106-5
Glycine-d Cambridge Isotope Laboratories Cat# DLM-1674-5
Kynurenic acid-d5 Cambridge Isotope Laboratories Cat#DLM-7374-PK
L-Citrulline-d4 Cambridge Isotope Laboratories Cat#DLM-6039-0.01
L-Threonine-(13C4,15N) Cambridge Isotope Laboratories Cat#CNLM-587-0.1
L-Tyrosine-d7 Cambridge Isotope Laboratories Cat#DLM-589-0.05
P-cresol sulfate-d7 Cambridge Isotope Laboratories Cat#DLM-9786-0.01
Sarcosine-d3 Cambridge Isotope Laboratories Cat#DLM-6874-0.1
Trans-4-hydroxy-L-proline-d3 Cambridge Isotope Laboratories Cat#DLM-9778-PK
Uric acid-(13C; 15N3) Cambridge Isotope Laboratories Cat#CNLM-10617-0.001
Critical Commercial Assays
Invitrogen total exosome isolation kit Thermo fisher Cat#4478360
Pierce BCA protein assay kit Thermo fisher Cat#23225
BD Multitest CD3/CD8/CD45/CD4 BD Bioscience Cat#340499
BD Trucount Tubes BD Bioscience Cat#340334
Deposited Data
Lipidomics and metabolomics raw datasets This paper http://dx.doi.org/10.17632/6z5zkzb3hv.3
Software and Algorithms
RX64-3.6.1 R Foundation for Statistical Computing https://www.r-project.org/
Analyst 1.6.3 Sciex https://sciex.com/products/software/analyst-software
MarkerView 1.3 Sciex https://sciex.com/products/software/markerview-software
PeakView 2.2 Sciex https://sciex.com/products/software/peakview-software
Other
K2 EDTA tube (10 ml) BD Vacutainer Cat# 367525
Luna Silica 3 μm column Phenomenex Cat#00F-4162-b0
Kinetex-C18 2.6 μm column Phenomenex Cat#00D-4462-e0
ACQUITY UPLC HSS T3 1.8 μm column Waters Cat#186004680

Resource Availability

Lead Contact
Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Guanghou Shui (ghshui@genetics.ac.cn).

Materials Availability
This study did not generate new unique reagents.

Data and Code Availability
These data are available at the Elsevier’s open research data repository website, Mendeley Data (https://data.mendeley.com). The data can be accessed directly via the dataset DOI: http://dx.doi.org/10.17632/6z5zkzb3hv.3.

Experimental Model and Subject Details

Study Participants and Data Collection
We retrospectively recruited a total of 50 patients with COVID-19 from January 22 to February 16, 2020, at the Fifth Medical Center of PLA General Hospital. All enrolled patients were confirmed to be positive for SARS-CoV-2 nucleic acid by real-time polymerase chain reaction. Twenty-six healthy individuals comprising doctors, nurses and researchers also stationed in the same hospital campus during the sample collection period were included as controls. The severity of COVID-19 was judged according to the guidelines for the diagnosis and management of COVID-19 patients (7th edition) by National Health Commission of China. Briefly, mild group (n = 18) included those present mild symptoms without pneumonia; moderate group (n = 19) present with fever, respiratory symptoms, and pneumonia; severe group (n = 13) included severe and critical ill cases. Severe cases are characterized by dyspnea, respiratory frequency ≥30/minute, blood oxygen saturation ≤93%, PaO2/FiO2 ratio < 300, and/or lung infiltrates greater than 50% within 24 to 48 h in pulmonary imaging. Critical ill cases refer to individuals that exhibited respiratory failure and required mechanical ventilation, as well as septic shock, and/or multiple organ dysfunction/failure that required monitoring and treatment in ICU. Our records indicated 2 out of 13 severe patients were on mechanical ventilation at the point of blood sample collection. Demographic, clinical, and laboratory radiological data were extracted from electronic medical records. Baseline characteristics and laboratory findings of COVID-19 patients and healthy controls were summarized in Tables 1 and S1–S3. The admission data of these patients were collected and checked independently by two physicians. The study was performed in accordance with the Declaration of Helsinki principle for ethical research. The study protocol was approved by Ethics Committee of the Fifth Medical Center of PLA General Hospital. Written informed consent was waived by the Ethics Committee of the designated hospital for emerging infectious disease.

Method Details

Plasma Collection and Metabolite Extraction
All blood samples used in this study were collected after an overnight fast. Peripheral blood samples from patients were collected within 24 h upon hospital admission. For patients admitted in the evening, blood samples were taken on the next morning (0500-0600) before breakfast (0700). For patients admitted during daytime, blood samples were collected only on the next morning (0500-0600) before breakfast (0700) (i.e., after one night in the hospital). Throughout the hospitalization period, patients were provided two standard meals per day, scheduled at 0700 and 1700, respectively. Blood was collected in BD Vacutainer (BD 367525). Plasma was separated by centrifugation at 2000 rpm for 10 min. Lipids and metabolites were extracted according to a modified version of the Bligh and Dyer’s protocol (Lu et al., 2019). Plasma (100 μL) for metabolomics and lipidomics analyses was inactivated via the addition of 750 μL of ice-cold chloroform: methanol (1:2) (v/v). Samples were vortexed for 15 s and then incubated for 1 h at 1500 rpm at 4°C. At the end of incubation, 250 μL of ice-cold chloroform and 350 μL of ice-cold MilliQ water were added. Samples were vortexed for 15 s and put on ice for 1 min. This step was repeated once. Samples were then centrifuged at 12 000 rpm for 5 min 4°C to induce phase separation. The lower organic phase was first extracted to a new tube. Then, another 450 μL of ice-cold chloroform was added to the remaining aqueous/methanol phase. Samples were vortexed briefly for 15 s and put on ice for 1 min, and centrifuged at 12 000 rpm for 5 min 4°C. The lower organic phase was extracted and pooled together with the first round organic extract. Double rounds of extraction ensured a better recovery and reduce variations across samples. The remaining aqueous/methanol phase was then centrifuged at 12 000 rpm for 5 min 4°C, and clean supernatant containing polar metabolites were extracted and transferred to new tube. The organic phase was dried in the SpeedVac under OH mode, while aqueous phase was dried under H2O mode. The dried metabolite extracts were shipped on dry ice to the designated laboratory for lipidomics and metabolomics analysis.

Exosome Isolation and Lipid Extraction
Exosomes were isolated from 100 μL of plasma using Invitrogen total exosome isolation kit (Thermofisher Scientific) according to the manufacturer’s protocol. The isolated exosome pellet was resuspended in 100 μL of ice-cold PBS and dispersed completely by repeatedly pipetting up and down. Following this, 750 μL of ice-cold chloroform: methanol (1:2) (v/v) was added to inactivate the samples. Lipid were then extracted in identical steps as described above for plasma samples, and organic extracts pooled from two rounds of extractions were used for lipidomics analysis. The remaining aqueous/methanol phase containing the extracted pellet was dried in the Speedvac under H2O mode. Proteins were extracted from the dried pellet using RIPA lysis buffer with protease inhibitor cocktail (Sigma-Aldrich), and total protein content was determined using Pierce BCA protein assay kit (Thermofisher Scientific) according to the manufacturer’s instructions.

Targeted Lipidomics
Prior to analysis, plasma lipid extracts were resuspended in 100 μL of chloroform: methanol 1:1 (v/v) spiked with appropriate concentrations of internal standards. All lipidomic analyses were carried out on an Exion UPLC coupled with a SCIEX QTRAP 6500 PLUS system as described previously, using an extensive, targeted library tailored for human serum lipidome that confers sufficient lipid coverage to render global lipid pathway analysis (Lam et al., 2014; Lu et al., 2019). All quantification experiments were conducted using internal standard calibration. In brief, polar lipids were separated on a Phenomenex Luna Silica 3 μm column (i.d. 150 × 2.0 mm) under the following chromatographic conditions: mobile phase A (chloroform:methanol:ammonium hydroxide, 89.5:10:0.5) and mobile phase B (chloroform:methanol: ammonium hydroxide: water, 55:39:0.5:5.5) at a flow rate of 270 μL/min and column oven temperature at 25°C. The gradient started with 5% of B and was held for 3 min, which was then increased to 40% of B over 9 min, and was held at 40% for 4 min before further increasing to 70% B over 5 min. The gradient was maintained at 70% B for 15 min before returning to 5% B over 3 min, and was finally equilibrated for 6 min. Individual polar lipid species were quantified by referencing to spiked internal standards of the same lipid class including PC-14:0/14:0, d31-PC16:0/18:1, PE14:0/14:0, d31-PE-16:0/18:1, d31-PS-16:0/18:1, PA-17:0/17:0, PG-14:0/14:0, d31-PG-16:0/18:1, C14:0-BMP, d31-PI-16:0/18:1, SM-d18:1/12:0, LPC-17:0, LPE-17:1, LPI-17:1, LPA-17:0, LPS-17:1, S1P-d17:1, Cer-d18:1/d7-15:0, GluCer d18:1/8:0, GalCer d18:1/8:0 obtained from Avanti Polar Lipids (AL, USA) and PI-8:0/8:0 from Echelon Biosciences, Inc. (UT, USA). d3-GM3 d18:1/18:0 and d3-LacCer d18:1/16:0 were from Matreya LLC (PA, USA). Glycerol lipids including diacylglycerols (DAGs) and triacylglycerols (TAGs) were quantified using a modified version of reverse phase LC/MRM. Separation of neutral lipids were achieved on a Phenomenex Kinetex-C18 2.6 μm column (i.d. 4.6x100 mm) using an isocratic mobile phase containing chloroform:methanol:0.1 M ammonium acetate 100:100:4 (v/v/v) at a flow rate of 300 μL for 10 min. Levels of short-, medium-, and long-chain TAGs were calculated by referencing to spiked internal standards of TAG(14:0)3-d5, TAG(16:0)3-d5 and TAG(18:0)3-d5 obtained from CDN isotopes (Quebec, Canada), respectively. DAGs were quantified using d5-DAG16:0/16:0 and d5-DAG18:1/18:1 as internal standards from Avanti Polar Lipids (Shui et al., 2010). Free cholesterols and cholesteryl esters were analyzed as described previously with d6-cholesterol and d6-CE18:0 cholesteryl ester (CE) (CDN isotopes) as internal standards (Shui et al., 2011). Lipid levels were expressed in nanomoles per L (nmol/L) for plasma, and in nanomoles lipids per g of total protein (nmol/g) for exosomes.

Untargeted Metabolomics
Prior to analysis, aqueous extracts were resuspended in 100 μL of 2% acetonitrile in water. Chromatographic separation was performed on a reversed-phase ACQUITY UPLC HSS T3 1.8 μm column (i.d. 3.0 × 100 mm) (Waters) using an UPLC system (Agilent 1290 Infinity II; Agilent Technologies) as described previously (Tian et al., 2020). MS detection was performed using high-resolution time-of-flight (TOF) mass spectrometry (5600 Triple TOF Plus, Sciex) equipped with an ESI source (Yuan et al., 2012). Data were acquired in TOF full scan method with positive and negative ion modes, respectively. Information-dependent acquisition methods were used for MS/MS analyses of metabolome. The collision energy was set at 35 ± 15 eV. Metabolite identification was compared with standard references, HMDB (https://hmdb.ca/), METLIN (https://metlin.scripps.edu), and literature searches. A total of 45 isotopically-labeled internal standards (IS), purchased from Cambridge Isotope Laboratories, were spiked into the samples for metabolite quantitation, including L-Phenylalanine-d8, L-Tryptophan-d8, L-Isoleucine-d10, L-Asparagine-13C4, L-Methionine-d3, L-Valine-d8, L-Proline-d7, L-Alanine-d7, DL-Serine-d3, DL-Glutamic acid-d5, L-Aspartic acid-d3, L-Arginine-d7, L-Glutamine-d5, L-Lysine-d9, L-Histidine-d5, Taurine-d2, Betaine-d11, Urea-(13C,15N2), L-lactate-13C3, Trimethylamine N-oxide-d9, Choline-d13, Malic acid-d3, Citric acid-d4, Succinic acid-d4, Fumaric acid-d4, Hypoxanthine-d3, Xanthine-15N2, Thymidine (13C10,15N2), Inosine-15N4, Cytidine-13C5, Uridine-d2, Methylsuccinic acid-d6, Benzoic acid-d5, Creatine-d3, Creatinine-d3, Glutaric acid-d4, Glycine-d2, Kynurenic acid-d5, L-Citrulline-d4, L-Threonine-(13C4,15N), L-Tyrosine-d7, P-cresol sulfate-d7, Sarcosine-d3, Trans-4-hydroxy-L-proline-d3, Uric acid-(13C; 15N3). Metabolite levels were normalized according to the following rules (1) ISs were applied to correct peak areas of their corresponding metabolites; (2) When (1) was not feasible due to unavailability of commercial standards, peak areas were corrected with IS of metabolites of the same class, comparable peak intensities, and/or proximity in retention times; (3) Results from Step (2) were evaluated based on relative standard deviation (RSD) values of each metabolite before and after IS correction. Corrected peak areas were adopted if their corresponding RSDs were smaller than that of original areas in quality control samples. For simplicity, all metabolite levels were labeled as “intensity.”

Quantification and Statistical Analysis

Metabolite Panel for Identifying COVID-19
To generate a plasma metabolite panel for differentiating COVID-19 patients from healthy individuals, variables with p < 0.05 between healthy controls and COVIDP19 patients after adjustment for age, sex and BMI were sieved out to form a starting pool. From this pool, a starting variable with lowest p value was added to Set 1, and remaining variables from the starting pool significantly correlated (p < 0.05) with this starting variable were added together to form Set 1. The process then was repeated in an iterative process using starting variable with the second lowest p value, and so on, finally generating a total of ten established sets. Representative metabolite from each established set was chosen based on (1) smallest p value and (2) reported biological function through a PubMed search. The selection process finally created a panel of ten plasma metabolites, and its performance was evaluated in a logistic regression model with leave-one-out (LOO) cross-validation, which distinguished between COVID-19 patients and healthy controls with an area under curve (AUC) = 0.955.

Logistic Regression Analysis
Logistic regression model with covariates BMI, age and sex were built with each variable (lipid/metabolite) to search for significant variables that can predict the conditions (i.e., healthy control, mild, moderate, severe COVID-19) of subjects. The p value of the variable estimate in the model was extracted and those with false discovery rate (fdr) smaller than 0.05 were shortlisted. Forest plots of these significant variables were constructed and the boxplots of individual significant variables were generated. A list of 61 lipids and 89 metabolites were found to be significant (fdr < 0.05) in their respective logistic regression models. Forest plots illustrate the magnitude of estimates on signed log x axis, with indicator of significance of the estimate in the model, with ∗∗∗ representing p < 0.001, ∗∗ representing p < 0.01 and ∗ representing p < 0.05. For non-significant species, the estimates were plotted as zeros.

Spearman Correlation Analysis
Spearman correlation between variables and clinical indicators were performed with data from COVID-19 patients. For each pair of variable and clinical indicator, samples with missing clinical data were omitted from the calculations. Correlation plots for all variables grouped by metabolite class were presented. Only correlations with p < 0.05 were indicated with colored circles. Negative correlations were shown in red and positive correlations were shown in blue, with sizes of circles representing the magnitude of the correlations.

Differential Correlation Analysis
MEGENA R package was used to build correlation networks from differentially correlated lipid pairs in mild COVID-19 relative to healthy controls to reveal changes in lipid co-regulation upon early infection. Differential correlation was calculated using R package DGCA. Only lipid pairs with differential correlation (empirical p < 0.05) were included for analyses.