Abstract
Cytokine storms and hyperinflammation, potentially controlled by glucocorticoids, occur in COVID-19; the roles of lipid mediators and acetylcholine (ACh) and how glucocorticoid therapy affects their release in Covid-19 remain unclear. Blood and bronchoalveolar lavage (BAL) samples from SARS-CoV-2- and non-SARS-CoV-2- infected subjects were collected for metabolomic/lipidomic, cytokines, soluble CD14 (sCD14), and ACh, and CD14 and CD36-expressing monocyte/macrophage subpopulation analyses. Transcriptome reanalysis of pulmonary biopsies was performed by assessing coexpression, differential expression, and biological networks. Correlations of lipid mediators, sCD14, and ACh with glucocorticoid treatment were evaluated. This study enrolled 190 participants with Covid-19 at different disease stages, 13 hospitalized non-Covid-19 patients, and 39 healthy-participants. SARS-CoV-2 infection increased blood levels of arachidonic acid (AA), 5-HETE, 11-HETE, sCD14, and ACh but decreased monocyte CD14 and CD36 expression. 5-HETE, 11-HETE, cytokines, ACh, and neutrophils were higher in BAL than in circulation (fold-change for 5-HETE 389.0; 11-HETE 13.6; ACh 18.7, neutrophil 177.5, respectively). Only AA was higher in circulation than in BAL samples (fold-change 7.7). Results were considered significant at P<0.05, 95%CI. Transcriptome data revealed a unique gene expression profile associated with AA, 5-HETE, 11-HETE, ACh, and their receptors in Covid-19. Glucocorticoid treatment in severe/critical cases lowered ACh without impacting disease outcome. We first report that pulmonary inflammation and the worst outcomes in Covid-19 are associated with high levels of ACh and lipid mediators. Glucocorticoid therapy only reduced ACh, and we suggest that treatment may be started early, in combination with AA metabolism inhibitors, to better benefit severe/critical patients.
Introduction
Individuals Covid-19 may present asymptomatically or with manifestations ranging from acute respiratory distress syndrome to systemic hyperinflammation and organ failure, events attributed to cytokine storms1. Free polyunsaturated fatty acids, such as AA and derivative eicosanoids, regulate inflammation2, 3, yet their role in Covid-19 has not been well investigated.
ACh, which is released by nerves4, leukocytes5, and airway epithelial cells6, regulates metabolism7, cardiac function8, airway inflammation9, and cytokine production10, all of which occur in Covid-19. It is known that eicosanoids stimulate ACh release3, but crosstalk between cholinergic and lipid mediator pathways in Covid-19 still need to be clarified.
In this study, levels of lipid mediators, ACh, and other inflammatory markers in blood and BAL from patients with Covid-19 who were treated or not with glucocorticoids were compared to those of non-Covid-19 and healthy-participants. Moreover, lung biopsy transcriptome reanalysis data from Covid-19 and non-Covid-19 patients corroborated our findings.
Methods
Study design and blood collection
This observational, analytic, and transversal study was conducted from June to November 2020. All participants were over 16 years old and chosen according to the inclusion and exclusion criteria described in Table S1 and in the protocol, after providing signed consent. Blood samples collected from patients positive for Covid-19 (n=190) were analyzed by RT-qPCR (Biomol OneStep/Covid-19 kit; Institute of Molecular Biology of Paraná - IBMP Curitiba/PR, Brazil) using nasopharyngeal swabs and/or serological assays to detect IgM/IgG/IgA (SARS-CoV-2® antibody test; Guangzhou Wondfo Biotech, China). Samples obtained from a cohort of SARS-CoV-2-negative healthy participants were used as controls (n=39). Participants positive for Covid-19 were categorised as asymptomatic-mild (n=43), moderate (n=44), severe (n=54), or critical (n=49). The criteria for the clinical classification of patients were defined at the time of sample collection, as shown in Table S1. Peripheral blood samples were obtained by venous puncture from patients upon their first admission and/or during the period of hospitalisation at two medical centres, Santa Casa de Misericordia de Ribeirão Preto and Hospital Sao Paulo at Ribeirão Preto, São Paulo State, Brazil. Blood samples from healthy controls and asymptomatic-mild non-hospitalized participants were collected either at the Centre of Scientific and Technological Development “Supera Park” (Ribeirão Preto, São Paulo State, Brazil) or in the home of patients receiving at-home care. The plasma was separated from whole blood samples and stored at −80°C. For lipidomic and metabolomic analyses, 250 µL of plasma was stored immediately in methanol (1:1 v/v). Lipidomic and metabolomic analyses were performed by mass spectrometry (LC-MS/MS), while a cytokines, sCD14, and ACh were quantified using CBA flex Kit (flow cytometer assay) or commercial ELISA. The expression levels of CD14, CD36, CD16, and HLA-DR in cells were evaluated by flow cytometry following the gate strategy (Figure S2).
Ethical considerations
All participants provided written consent in accordance with the regulations of the Conselho Nacional de Pesquisa em Humanos (CONEP) and the Human Ethical Committee from Faculdade de Ciências Farmacêuticas de Ribeirão Preto (CEP-FCFRP-USP). The research protocol was approved and received the certificate of Presentation and Ethical Appreciation (CAAE: 30525920.7.0000.5403). The sample size was determined by the convenience of sampling, availability at partner hospitals, agreement to participate, and the pandemic conditions within the local community (more information in the Protocol).
Bronchoalveolar lavage fluid (BAL) collection and processing
BAL fluids were collected from hospitalised Covid-19 patients at the severe or critical stages of disease (n=32) to assess their lung immune responses. Control samples were obtained from hospitalised intubated donors negative for SARS-CoV-2 (n=13) (as certified by SARS-CoV-2-negative PCR), referred to as non-Covid-19 patients, who were intubated because of the following primary conditions: bacterial pneumonia, abdominal septic shock associated with respiratory distress syndrome, pulmonary atelectasis due to phrenic nerve damage, or pulmonary tuberculosis. BAL fluid was collected as previously described11, using a siliconized polyvinylchloride catheter (Mark Med, Porto Alegre, Brazil) with a closed Trach Care endotracheal suction system (Bioteque Corporation, Chirurgic Fernandes Ltd., Santana Parnaíba, Brazil) and sterile 120 mL polypropylene flask (Biomeg-Biotec Hospital Products Ltd., Mairiporã, Brazil) under aseptic conditions. Approximately 5–10 mL of bronchoalveolar fluid was obtained and placed on ice for processing within 4 h. The BAL fluids were placed into 15-mL polypropylene collection tubes and received half volume of their volume of phosphate buffered saline (PBS) 0.1 M (2:1 v/v) in relation to the total volume of each sample. After centrifugation (700 × g, 10 min), the supernatants of the BAL fluid were recovered and stored at –80°C. For lipidomic and metabolomic analyses, 250 µL of these supernatants were stored immediately in methanol (1:1 v/v). Subsequently, the remaining BAL fluid was diluted in 10 mL of PBS and gently filtered through a 100-µm cell strainer (Costar, Corning, NY, USA) using a syringe plunger. The resulting material was used for cytokine and acetylcholine (ACh) quantification. The BAL fluids were centrifuged (700 × g, 10 min) and the red blood cells were lysed using 1 mL of ammonium chloride (NH4Cl) buffer 0.16 M for 5 min. The remaining airway cells were washed with 10 mL of PBS, resuspended in PBS–2% heat-inactivated foetal calf serum, and counted with Trypan blue using an automated cell counter (Countess, Thermo Fisher Scientific, Waltham, MA, USA). The leukocyte numbers were adjusted to 1 × 109 cells/L for differential counts and 1 × 106 cells/mL for flow cytometry analysis. All procedures were performed in a Level 3 Biosafety Facility (Departamento de Bioquímica e Imunologia, Faculdade de Medicina de Ribeirão Preto, Universidade de São Paulo).
Data collection
The electronic medical records of each patient were carefully reviewed. Data included sociodemographic information, comorbidities, medical history, clinical symptoms, routine laboratory tests, immunological tests, chest computed tomography (CT) scans, clinical interventions, and outcomes (more information in the Protocol). The information was documented on a standardised record form, as indicated in Tables S1, S2, and S3. Data collection of laboratory results included first-time examinations within 24 h of admission, defined as the primary endpoint. The secondary endpoint was clinical outcome (death or recovery).
Clinical laboratory collection
For hospitalised patients, blood examinations were performed by clinical analysis laboratories at their respective hospitals. Blood examinations of healthy participants and non-hospitalized patients were performed at Serviço de Análises Clínicas (SAC), Departamento de Análises Clínicas, Toxicológicas e Bromatológicas of the Faculdade de Ciências Farmacêuticas de Ribeirão Preto, Universidade de São Paulo, Ribeirão Preto, São Paulo, Brazil. The blood samples were used to measure for liver and kidney function, myocardial enzyme spectrum, coagulation factors, red blood cells, haemoglobin, platelets, and total and differential leukocytes using automated equipment. Similarly, the absolute numbers of leukocytes in the BAL fluid were determined in a Neubauer Chamber with Turkey solution. For the counts of differential leukocytes in the BAL, 100 µL of the fluid was added to cytospin immediately after collection to avoid any interference on cell morphology. Differential leukocyte counts were conducted using an average of 200 cells after staining with Fast Panoptic (LABORCLIN; Laboratory Products Ltd, Pinhais, Brazil) and examined under an optical microscope (Zeiss EM109; Carl Zeiss AG, Oberkochen, Germany) with a 100× objective (immersion oil) equipped with a Veleta CCD digital camera (Olympus Soft Imaging Solutions Gmbh, Germany) and ImageJ (1.45s) (National Institutes of Health, Rockville, MD, USA)12. Lymphocytes, neutrophils, eosinophils, and monocytes/macrophages were identified and morphologically characterised, and their lengths and widths were measured (100×).
High-performance liquid chromatography coupled with tandem Mass Spectrometry (LC-MS/MS) assay
Reagents
Eicosanoids, free fatty acids (AA, EPA, and DHA), and metabolites as molecular weight standards (MWS) and deuterated internal standards were purchased from Cayman Chemical Co. (Ann Arbor, MI, USA). HPLC-grade acetonitrile (ACN), methanol (MeOH), and isopropanol were purchased from Merck (Kenilworth, NJ, USA). Ultrapure deionised water (H2O) was obtained using a Milli-Q water purification system (Merck-Millipore, Kenilworth, NJ, USA). Acetic acid (CH3COOH) and ammonium hydroxide (NH4OH) were obtained from Sigma Aldrich (St. Louis, MO, USA).
Sample preparation and extraction
The plasma (250 μL) in EDTA-containing tubes (Vacutainer® EDTA K2; BD Diagnostics, Franklin Lakes, NJ, USA) and BAL (250 μL) samples were stored in MeOH (1:1, v/v) at ‒80°C. Three additional volumes of ice-cold absolute MeOH were added to each sample overnight at ‒20°C for protein denaturation and after lipid solid-phase extraction (SPE). To each sample, 10 μL of internal standard (IS) solution was added, centrifuged at 800 × g for 10 min at 4°C. The resulting supernatants were collected and diluted with deionised water (ultrapure water; Merck-Millipore, Kenilworth, NJ, USA) to obtain a MeOH concentration of 10% (v/v). In the SPE extractions, a Hypersep C18-500 mg column (3 mL) (Thermo Scientific-Bellefonte, PA, USA) equipped with an extraction manifold collector (Waters-Milford, MA, USA) was used. The diluted samples were loaded into the pre-equilibrated column and washing using 2 mL of MeOH and H2O containing 0.1% acetic acid, respectively. Then, the cartridges were flushed with 4 mL of H2O containing 0.1% acetic acid to remove hydrophilic impurities. The lipids that had been adsorbed on the SPE sorbent were eluted with 1 mL of MeOH containing 0.1% acetic acid. The eluates solvent was removed in vacuum (Concentrator Plus, Eppendorf, Germany) at room temperature and reconstituted in 50 μL of MeOH/H2O (7:3, v/v) for LC-MS/MS analysis.
LC-MS/MS analysis and lipids data processing
Liquid chromatography was performed using an Ascentis Express C18 column (Supelco, St. Louis, MO, USA) with 100 × 4.6 mm and a particle size of 2.7 μm in a high- performance liquid chromatography (HPLC) system (Nexera X2; Shimadzu, Kyoto, Japan). Then, 20 μL of extracted sample was injected into the HPLC column. Elution was carried out under a binary gradient system consisting of Phase A, comprised of H2O, ACN, and acetic acid (69.98:30:0.02, v/v/v) at pH 5.8 (adjusted with NH4OH), and Phase B, comprised of ACN and isopropanol (70:30, v/v). Gradient elution was performed for 25 min at a flow rate of 0.5 mL/min. The gradient conditions were as follows: 0 to 2 min, 0% B; 2 to 5 min, 15% B; 5 to 8 min, 20% B; 8 to 11 min, 35% B; 11 to 15 min, 70% B; and 15 to 19 min, 100% B. At 19 min, the gradient was returned to the initial condition of 0% B, and the column was re-equilibrated until 25 min. During analysis, the column samples were maintained at 25°C and 4°C in the auto-sampler. The HPLC system was directly connected to a TripleTOF 5600+ mass spectrometer (SCIEX-Foster, CA, USA). An electrospray ionisation source (ESI) in negative ion mode was used for high-resolution multiple-reaction monitoring (MRMHR) scanning. An atmospheric-pressure chemical ionisation probe (APCI) was used for external calibrations of the calibrated delivery system (CDS). Automatic mass calibration (<2 ppm) was performed periodically after each of the five sample injections using APCI Negative Calibration Solution (Sciex-Foster, CA, USA) injected via direct infusion at a flow rate of 300 μL/min. Additional instrumental parameters were as follows: nebuliser gas (GS1), 50 psi; turbo gas (GS2), 50 psi; curtain gas (CUR), 25 psi; electrospray voltage (ISVF), ‒4.0 kV; temperature of the turbo ion spray source, 550°C. The dwell time was 10 ms, and a mass resolution of 35,000 was achieved at m/z 400. Data acquisition was performed using AnalystTM software (SCIEX-Foster, CA, USA). Qualitative identification of the lipid species was performed using PeakViewTM (SCIEX-Foster, CA, USA). MultiQuantTM (SCIEX-Foster, CA, USA) was used for the quantitative analysis, which allows the normalisation of the peak intensities of individual molecular ions using an internal standard for each class of lipid. The quantification of each compound was performed using internal standards and calibration curves, and the specific mass transitions of each lipid were determined according to our previously published method13. The final concentration of lipids was normalised by the initial volume of plasma or BAL fluid (ng/mL).
Metabolomics analysis
Metabolite was extracted and samples were transferred to autosampler vials for LC–MS analysis using TripleTOF5600+ Mass Spectrometer (Sciex-Foster, CA, USA) coupled to an ultra-high-performance liquid chromatography (UHPLC) system (Nexera X2; Shimadzu, Kyoto, Japan). Reverse-phase chromatography was performed similarly to lipids analyses above. Mass spectral data were acquired with negative electrospray ionisation, and the full scan of mass-to-charge ratio (m/z) ranged from 100 to 1500. Proteowizard software14 was used to convert the wiff files into mz XML files. Peak peaking, noise filtering, retention time, m/z alignment, and feature quantification were performed using apLCMS15. Three parameters were used to define a metabolite feature: mass-to-charge ratio (m/z), retention time (min), and intensity values. Data were log2 transformed and only features detected in at least 50% of samples from one group were used in further analyses. Missing values were imputed using half the mean of the feature across all samples. Mummichog (version 2) was used for metabolic pathway enrichment analysis (mass accuracy under 10 ppm)16.
Acetylcholine measurement
ACh was measured in heparinized plasma (SST® Gel Advance®; BD Diagnostics, Franklin Lakes, NJ, USA) and in BAL using a commercially available immunofluorescence kit (ab65345; Abcam, Cambridge, UK) according to the manufacturer’s instructions. Briefly, ACh was converted to choline by adding the enzyme acetylcholinesterase to the reaction, which allows for total and free-choline measurement. The amount of ACh present in the samples was calculated by subtracting the free choline from the total choline. The products formed in the assay react with the choline probe and can be measured by fluorescence with excitation and emission wavelengths of 535 and 587 nm, respectively (Paradigm Plate Reader; SpectraMax, San Diego, CA, USA). The concentration of ACh was analysed using SoftMax® software (SpectraMax, Molecular Devices, Sunnyvale, CA, USA), expressed as pmol.mL-1.
Soluble CD14 (sCD14) measurement
Samples from heparinized plasma (SST® Gel Advance®; BD Biosciences, Franklin Lakes, NJ, USA) were placed in 96-well plates. The concentration of sCD14 was determined using an ELISA kit (DY383; R&D Systems, Minneapolis, MN, USA), following the manufacturer’s instructions, expressed as pg.mL-1.
Flow Cytometry
Uncoagulated blood samples in EDTA-containing tubes (Vacutainer® EDTA K2; BD Biosciences) were processed for flow cytometry analysis of circulating leukocytes. Whole blood (1 mL) was separated and red blood cells were lysed using RBC lysis buffer (Roche Diagnostics GmbH, Mannheim, GR). Leukocytes were washed in PBS containing 5% foetal bovine serum (FBS) (Gibco™, USA), centrifuged, and resuspended in Hank’s balanced salt solution (Sigma-Aldrich, Merck, Darmstadt, Germany) containing 5% FBS, followed by surface antigen staining. Similarly, cells obtained from BAL fluid were processed for flow cytometry assays. Briefly, cells were stained with Fixable Viability Stain 620 (1:1000) (BD Biosciences) and incubated with monoclonal antibodies specific for CD14 (1:100) (M5E2; Biolegend), HLA-DR (1:100) (G46-6; BD Biosciences), CD16 (1:100) (3G8; Biolegend), and CD36 (1:100) (CB38, BD Biosciences) for 30 min at 4°C. Stained cells were washed and fixed with BD Cytofix™ Fixation Buffer (554655; BD Biosciences, San Diego, CA, USA). Data acquisition was performed using a LSR-Fortessa™ flow cytometer (BD Biosciences, San Jose, CA, USA) and FACS-Diva software (version 8.0.1) (BD Biosciences, Franklin Lakes, NJ, USA). For the analysis, 300,000 events were acquired for each sample. Data were evaluated using FlowJo® software (version 10.7.0) (Tree Star, Ashland, OR, USA) to calculate the cell frequency, dimensionality reduction, and visualisation using t-distributed stochastic neighbour embedding. Gate strategy performed as described before 17, as shown in Figure S2.
Cytokine Measurements
The cytokines interleukin (IL)-6, IL-8, IL-1ß, IL-10, and tumour necrosis factor (TNF) were quantified in heparinized plasma and BAL fluid samples using a BD Cytometric Bead Array (CBA) Human Inflammatory Kit (BD Biosciences, San Jose, CA, USA), according to manufacturer’s instructions. Briefly, after sample processing, the cytokine beads were counted using a flow cytometer (FACS Canto TM II; BD Biosciences, San Diego, CA, USA), and analyses were performed using FCAP Array (3.0) software (BD Biosciences, San Jose, CA, USA). The concentrations of cytokines were expressed as pg.mL-1.
Re-analysis of transcriptome data from lung biopsies of patients with Covid-19
To gain a better understanding of the correlation between the altered concentrations of ACh, AA, and AA-metabolites detected in the plasma and BAL fluid of severe/critical Covid-19 patients, we performed a new analysis by re-using a previously published transcriptome open dataset 18, deposited in the Gene Expression Omnibus repository under accession no. GSE150316 19. We used transcriptome data from lung samples (n=46) from patients with Covid-19 (n=15), seven of which displayed a low viral load and eight a high viral load, and non-Covid-19 patients (n=5) with other pulmonary illnesses (negative control). Patients with a high viral load had meantime periods of hospital stays (3.6±2.2 days) and duration of illness (7.2±3.02 days) shorter than the patients with low viral load (14±7.9 and 19±4.9 days, respectively), as described by the authors of the public data source 18. Hence, for analysis purposes, all patient samples were grouped into four classifications: Covid-19 (CV), Covid-19 low viral load (CVL), Covid-19 high viral load (CVH), and non-Covid-19 (NCV). The strategy for reanalysing the transcriptome was implemented according to three consecutive steps: (i) co-expression analysis, (ii) differential expression analysis, and (iii) biological network construction. Initially, for the co-expression study, normalised transcriptome data in log2 of reads per million (RPM) were filtered by excluding non-zero counts in at least 20% of the samples. Next, the selected genes were explored in the R package Co-Expression Modules identification Tool (CEMITool) 20, using a p-value of 0.05 as the threshold for filtering. Then, the co-expression modules were analysed for the occurrence of ACh and AA genes list obtained from the Reactome pathways 21, as well as the Covid-19-related genes obtained from the literature (Supplementary Appendix I). Next, differential gene expression between samples from the lung biopsy transcriptome (CV, NCV, CVL, and CVH) was measured using the DESeq2 package 22, with p-values adjusted using the Benjamini and Hochberg method 23. The list of differentially express genes (DEGs) generated for all comparisons was filtered from the genes listed in Supplementary Appendix I, considering the values of log2 of fold-change (FC) greater than 1 (|log2(FC)|>1) and adjusted p<0.05. Finally, a first-order biological network was constructed using co-expression module(s) containing genes associated with the ACh and AA pathways to characterise the interplay between these mediators in combination with Covid-19 severity markers, as well as to identify relevant DEGs and hub genes in this network, using the BioGRID repository 24. The networks were constructed, analysed, and graphically represented using the R packages igraph 25, Intergraph 26, and ggnetwork 27. Due to the substantial influence of glucocorticoid treatment on the levels of some mediators, we measured the sensitivity of genes from the differential expression analysis between CV samples from patients who underwent treatment (CTC, three patients and ten samples) and patients who were not treated (NCTC, 12 patients and 36 samples), as previously described 18.
Statistical Analysis
Two-tailed tests were used for the statistical analysis, with a significance value of p <0.05 and a confidence interval of 95%. The data were evaluated for a normal distribution using the Kolmogorov–Smirnov test. The parametric data were analysed using unpaired t-tests (for two groups) or one-way ANOVA followed by Tukey’s multiple comparison tests for three or more groups simultaneously. For data that did not display a Gaussian distribution, Mann-Whitney (for two-group comparisons) or Kruskal-Wallis testes were used, followed by Dunn’s post-tests for analysis among three or more groups. The cytokine network data in patients with Covid-19 were analysed using significant Spearman’s correlations at p<0.05. Data were represented by connecting edges to highlight positive strong (r ≥ 0.68; thick continuous line), moderate (0.36 ≥ r < 0.68; thinner continuous line), or weak (0 > r < 0.36; thin continuous line) and negative strong (r ≤ ‒0.68; thick dashed line), moderate (‒0.68 > r ≤ ‒0.36; thinner dashed line), or weak (-0.36 < r > 0; thin dashed line), as proposed previously 28, 29. The absence of a line indicates the non-existence of the relationship. The Venn diagrams were elaborated using the online tool Draw Venn Diagram (http://bioinformatics.psb.ugent.be/webtools/Venn/). The results were tabulated using GraphPad Prism software (version 8.0) and the differences were considered statistically significant at p<0.05. See the Additional Statistical Report section for more information. Some of the confounding variables associated with Covid-19 (age, sex, obesity, hypertension, and diabetes mellitus) were analysed for their potential impacts on the main analytical procedures of this study, such as ACh, AA, 5-HETE, and 11- HETE measurements of the plasma (healthy, asymptomatic-to-mild, moderate, severe, and critical patients) and BAL (severe and critical patients) samples. This analysis was performed using the Kruskal-Wallis, Mann-Whitney, Spearman’s correlation, or Chi-square (χ2) tests (Table S7-S9).
Results
Study Population
This study enrolled 39 healthy-participants, 13 hospitalized non-Covid-19, and 190 Covid-19 patients aged 16-96 years from April to November 2020. The 190 Covid-19 patients were categorized as having asymptomatic-to-mild (n=43), moderate (n=44), severe (n=54), or critical (n=49) disease (Table S2).
Covid-19 Modifies Circulating Soluble Mediators and Cell Populations
To determine whether SARS-CoV-2 infection alters the metabolism of lipid mediators, we used high-resolution sensitive mass spectrometry to perform targeted eicosanoid analysis and nontargeted metabolomics using plasma from healthy-participants and Covid-19 patients. In total, 8,791 metabolite features were present in at least 50% of all samples, and the relative abundance of 595 metabolite features (FDR adjusted P<0.05) was altered in the groups studied (Figure 1A). Two-way hierarchical clustering based on these significant metabolite features resulted in three clear clusters: one for severe/critical Covid-19 patients, one for healthy-participants and one for asymptomatic-to-mild and moderate Covid-19 (Figure S1A). Pathway analysis revealed the top significant metabolic pathways to be enriched in features involved in fatty acid biosynthesis, metabolism, activation, and oxidation (Figure 1B). Compared to healthy-participants, tentative metabolite annotations suggested an increased abundance of fatty acids (FFAs), such as linoleic acid, tetradecanoate, dodecanoate and AA, in COVID-19 (Figure 4C-F). Among the identified lipids, AA was the most abundant, and its levels correlated with the severity of Covid-19 (Figure 1F). Linoleic acid can be metabolized to AA, which in turn is a substrate for eicosanoids, such as 5-hydroxyeicosatetraenoic acid (5-HETE) and 11-hydroxyeicosatetraenoic acid (11-HETE) (Figure 1G, 1H); both molecules with function in the immune response. Overall, the increased plasma levels of AA in Covid-19 indicate that it predicts disease severity.
As eicosanoids induce cell recruitment and regulate immune responses, we next determined the profile of immune cells and soluble mediators in whole blood and plasma of patients with Covid-19 and healthy-participants. According to whole-blood analysis, absolute leukocyte and neutrophil counts (Figure 2A, 2B) were significantly higher but lymphocyte counts (Figure 2C) significantly lower in patients with severe/critical disease than in those with moderate disease. Eosinophil (Figure 2D) and basophil counts (Figure 2E) were reduced severe disease compared to asymptomatic and moderate disease.
Although no differences in total monocyte counts among the groups (Figure 2F) were observed based on CD16, CD14 and HLA-DR, expression of membrane CD14 was reduced in all SARS-CoV-2-infected patients compared to healthy-participants (Figure 2G; Figure S2, S3). In parallel, CD36 expression was decreased in monocytes from patients with severe/critical disease (Figure 2H), but sCD14 was increased only in plasma from critical patients (Figure 2I). We did not detect differences in classic or non-classic monocytes, but the percentage of intermediate monocytes was decreased in all SARS-CoV-2-infected participants compared to healthy-participants (Figure 2J, 2K, 2L). Based on plasma cytokine level analysis, patients with moderate, severe, and critical disease share a Covid-19 cytokine profile defined by increased IL-8, IL-6, and IL-10 (Figure 2M, 2N, 2Q) levels, with unaltered IL-1β and TNF levels (Figure 2O, 2P).
Covid-19 Induces Strong Lung Responses
We performed measurements of BAL from hospitalized Covid-19 and non-Covid-19 patients. Changes in lung metabolomics induced by SARS-CoV-2 infection (Figure 3A; Figure S1B) included alterations in sphingolipids, beta oxidation of trihydroxyprostanoil-CoA, biosynthesis and metabolism of steroidal hormones, vitamin D3, and glycerophospholipids (Figure 3B). We also evaluated AA and its metabolites. Despite no differences in AA, levels of 5-HETE and 11-HETE in BAL were significantly higher in Covid-19 than in non-Covid-19 patients, though other metabolites did not differ between these groups (Figure 3C). When assessing leukocytes in BAL between the patient groups, we found no differences in total or differential counts, with the exception of lymphocyte numbers (Figure 3D, 3E). In contrast to eicosanoids, cytokine profiles in Covid-19 and non-Covid-19 patients were similar (Figure 3F), suggesting that lipid mediators contribute to the pathophysiological processes induced by SARS-CoV-2. Interestingly, we observed a significant reduction in classical (Figure 3G) and intermediate (Figure 3H) monocytes in BAL from Covid-19 patients. In parallel, CD14 and CD36 expression in monocyte was lower in BAL of Covid-19 than in that of non-Covid-19 patients (Figure 3I, 3J).
Acetylcholine, Cytokines and Eicosanoids are Higher in the Lung Micro-Environment
We compared systemic and lung responses only in patients with severe/critical Covid-19 and found significantly higher levels of cytokines in BAL than in blood (Figure 4A). When comparing lipid mediator levels in either compartment from the same Covid-19 patients, we detected lower AA but higher levels of 5-HETE and 11-HETE in BAL than in blood (Figure 4B, 4C, 4D). Previous results from our group3 have shown that eicosanoids contribute to ACh release. Therefore, we next measured ACh in patients with Covid-19 and observed higher plasma levels of ACh in patients with Covid-19; in addition, ACh was elevated in patients with severe/critical disease compared with those with asymptomatic/moderate disease or healthy-participants (Figure 4E). Unexpectedly, in patients with severe/critical Covid-19, ACh levels in BAL were 10-fold higher than those in serum, and patients treated with glucocorticoids showed decreases in ACh in both compartments (Figure 4F, 4G). Interesting, neutrophil counts were higher in BAL, as these cells produce high levels of 5-HETE and IL-1β, both of which are mediators of ACh release30, 31 (Figure 4H). Correlation analysis was then performed to evaluate the relationship between eicosanoids, cytokines, sCD14 and ACh in patients with Covid-19. When comparing blood samples from all patients, we detected strong correlations between ACh versus IL-1β and moderate correlations between ACh versus AA. A substantial number of interactions between AA and its metabolites and between cytokines and eicosanoids were observed (Figure 4I; Figure S4A). Correlations among all parameters were also observed in BAL (Figure 4J; Figure S4B). To evaluate the benefit of glucocorticoid treatments and their relationship with eicosanoids, CD14 and ACh, we analysed the intersections in a Venn diagram of blood and BAL from severe/critical Covid-19. The results showed that treatment of Covid-19 with glucocorticoids did not have a significant influence on eicosanoid or sCD14 release in patients with more severe stages of disease; in critical patients, however, reductions of 44% and 65% in ACh levels in blood and BAL, respectively, were observed (Figure 4K-4Q; Figure S5). These findings suggest that the use of glucocorticoids has a positive effect on resolution of the inflammatory process because they reduce ACh release, which directly or indirectly stimulates cell recruitment and proinflammatory mediator release.
Altered Expression of Acetylcholine and Arachidonic Acid Pathway Genes in Lung Biopsies From Some Covid-19 Patients
In this study, we reanalysed the lung biopsy transcriptome from Covid-19 patients to evaluate expression of ACh and AA pathway genes (Supplementary Appendix I). These genes were coexpressed only in a single module (M1) that included genes related to the ACh release cycle, AA metabolism, cholinergic and eicosanoid receptors, and biomarkers of Covid-19 severity (Figure 5A; Table S4; Supplementary Appendix II). Furthermore, expression of nearly all genes was upregulated in some deceased Covid-19 patients with long hospital stays and low viral loads (Figure 5B; Figure S6A; Supplementary Appendix III). These differentially expressed genes populated the biological network and are likely under the action of some hubs (Figure 5C; Supplementary Appendix IV), such as oestrogen receptor II (ESR2) and albumin (ALB). We identified a unique proinflammatory gene expression profile in lung biopsies related to cholinergic and eicosanoid receptors (Figure 5H, 5I; Figure S6B, S6C). In combination with the altered levels of ACh, AA, and AA metabolites found in patient from our cohort, the transcriptome data reported herein strengthen the likelihood that these mediators contribute to Covid-19 severity (Figure 5D-5G). Interestingly, lung samples from Covid-19 patients with short hospital stays and high viral loads did not present this unique expression profile of ACh or AA pathway genes (Figure 5H, 5I; Figure S6B, S6C). Some Covid-19 patients treated versus not with glucocorticoid showed transcript expression of monoglyceride lipase (MGLL) and N-acylethanolamine acid amidase (NAAA) and up-regulation of fatty acid amide hydrolase (FAAH), which are involved in the production of AA from endocannabinoid (cases 3, 9, and 11; Figure 5B; Supplementary Appendix III). Besides, we detected other DEGs in biopsy samples from Covid-19 patients (glucocorticoid-treated versus non-treated) associated with AA, ACh, interferon pathways, and Covid-19 biomarkers (Supplementary Appendix III).
Discussion
A consensus is building around the fatal effects of SARS-CoV-2 infection, which is increasingly believed to cause death as a result of systemic hyperinflammation and multi-organ collapse32, secondary to systemic cytokine storms 33–35. However, few studies have compared pulmonary and systemic inflammation 36 or considered the contribution of eicosanoids and neurotransmitters to these effects. In our study, we hypothesised that, in Covid-19, pulmonary cells and leukocytes, in addition to cytokines, release eicosanoids and ACh, mediating local and systemic manifestations. In patients with severe/critical SARS-CoV-2 infection, we found that ACh, 5-HETE, 11-HETE, and cytokines were more abundant in the lung than under systemic conditions. In contrast, only the levels of AA were found to be higher in the circulation than in the BAL fluid. Interestingly, in patients with severe/critical disease who were treated with corticosteroids, only ACh was inhibited.
In our study, we compared the lung and systemic responses in association with the lung transcriptome and demonstrated a robust correlation between lipid mediators, neurotransmitters, and their receptors in SARS-CoV-2 infection. However, contrary to what has been suggested by previous studies 37, 38, only small amounts of eicosanoids were found in the plasma of patients with severe/critical disease, with significant differences observed only in AA, 5-HETE, and 11-HETE. Interestingly, the levels of 5-HETE and 11-HETE were found to be remarkably higher in BAL, as well as AA in plasma, suggesting that AA and its metabolites mediate responses to Covid-19. 5-HETE induces neutrophil recruitment 39, pulmonary oedema 40, and ACh release 30. It is released by human neutrophils, and its esterified form promotes IL-8 secretion 31. Unlike the esterified form 31, free-5-HETE did not inhibit NETs formation, a key event in Covid-1941. Remarkably, 11-HETE originating from monocytes/macrophages 42, endothelial cells 43, and platelets 44 is induced by hypoxia 45, IL-1β 46, contributes to ACh functions 43, and inhibits insulin release 47. The involvement of AA was confirmed by bioinformatic analyses that identified the expression or upregulation of genes related to AA metabolism and eicosanoid receptors in some lung biopsies. These included the OXER1 gene, which encodes a receptor for AA and 5-HETE 48 which mediates neutrophil activation/recruitment 49, 50. Our metabolomic analysis also showed an increase in the plasma AA and linoleic acid levels, similar to a plasma lipidome performed by Schwarz et al., suggesting a strong correlation between lipid mediators and Covid-19 severity 51. Accordingly, the ELOVL2 gene, which is involved in linoleic acid metabolism and AA synthesis 52, was upregulated in lung biopsies.
The RNA expression of ALOX5, an enzyme that participates in lipid mediator production, is upregulated in some immune cell types from severe Covid-19 patients 53, and has been detected in the lung biopsies of deceased Covid-19 patients 18. In contrast to our lung transcriptome re-analysis findings, the expression of cytochrome p450 (CYP) enzymes, which are also involved in lipid mediator generation, was not detected in the peripheral blood mononuclear cells (PBMC) transcriptome re-analysis from severe Covid-19 patients 53. One possible explanation for this contradiction is that Covid-19 is a heterogeneous illness composed of distinct tissue gene expression associated with Covid-19 severity and different immunopathological profiles in infected tissues 18, 54. This immunological heterogeneity could be related to two types of evolution of severe Covid-19 patients, one associated with a high viral load and susceptibility to SARS-CoV-2 infection, a shorter hospitalisation time, and exudative diffuse alveolar damage, and another associated with a low or undetectable viral load, a mixed lung histopathological profile, and a longer hospitalisation time associated with non-homeostatic pulmonary inflammation 18, 54, 55. In this context, we found that the expression levels of some lipid mediators and ACh and AA pathway genes varied in the plasma and BAL samples of severe/critical patients, which were found to be preferentially activated in the lungs of deceased Covid-19 patients with low viral loads, long hospitalisation times, and damaging lung inflammation.
Nevertheless, it remains unclear whether AA and linoleic acid contribute to host protection 56 and tissue damage, or whether they represent a viral escape mechanism. AA is known to directly interact with the virus, reducing its viability and inducing membrane disturbances, disfavouring SARS-CoV-2 entry 57–59. On the other hand, reduced plasma AA levels may be associated with lung injury and poor outcomes in Covid-19 infection60. Within this context, Shen and et al. found lower concentrations of AA in survivors of severe Covid-19 infection 37. However, in the present study, the levels of AA were found to be increased in the plasma of patients who died from severe Covid-19, suggesting that the potential benefits of AA are overcome by a higher production of its proinflammatory metabolites, 5-HETE and 11-HETE. Interestingly, in our cohort, AA, 5-HETE, and 11-HETE were not altered in the plasma or BAL fluid of patients with severe/critical Covid-19 infection who had received glucocorticoid treatment. This may be explained by the fact that AA, in addition to calcium-activated-PLA2 61 also originates from glucocorticoid-insensitive phospholipase 62, adipocyte destruction 63, linoleic acid 64, or the degradation of endocannabinoids by FAAH 65, a macrophage enzyme detected in SARS-CoV-2-infected patients 66. Notably, some CVL Covid-19 patients, regardless of whether they were treated or not with glucocorticoids, showed transcript expression of MGLL and NAAA, as well as an up-regulation of FAAH and all of the enzymes involved in the production of AA from endocannabinoid 67. In addition, we detected other DEGs in biopsy samples (glucocorticoid-treated versus non-treated) associated with AA, ACh, interferon pathways, and Covid-19 biomarkers. Glucocorticoids have been used widely to reduce the morbidity and mortality rates of Covid-19 patients, but have not been effective for all patients 68. The mortality rate for glucocorticoid-treated hospitalised Covid-19 patients from our cohort was higher than that reported in the RECOVERY trial and similar to that described in the CoDEX trial conducted in Brazil 69. High mortality rates could be associated with several factors, including a low mean PaO2:FiO2 ratio and overloaded public health systems in countries with limited resources, such as Brazil 69 and as well as glucocorticoid dose, initiation, and duration of therapy 70, 71. In addition, the precise threshold at which a patient should be treated with glucocorticoids, that is to avoid the manifestation of adverse effects associated with comorbidities and inefficient clearance of SARS-CoV-2, remains unclear 72. However, a delayed start of glucocorticoid therapy could result in a lack of response due to patients reaching a point of no return, as suggested previously by our research group with regards to scorpion poisoning, which triggers a sterile inflammatory process 3. On the other hand, considering that Covid-19 is a non-sterile hyperinflammatory disease, the administration of glucocorticoids is more effective after the initial phase in which hospitalised patients have low or undetectable viral loads 18, 68. Our data suggest that SARS-CoV-2 infection activates the glucocorticoid-insensitive release of AA metabolites associated with Covid-19 severity. Hence, lipid mediator production pathways could be important molecular targets for Covid-19 treatment, as suggested by other researchers 53. Accordingly, we suggested that, in order to improve the benefits of glucocorticoid therapy in hospitalized patients, who show lower or undetectable viral loads, patients should be treated as soon as possible in combination with AA metabolism inhibitors.
CD36 is expressed in several cells 73, 74, where it induces Ca++ mobilisation, cell signalling, LTB4 production, and cellular fatty acid uptake 75–78. The maintenance of high free-AA levels in patients with severe/critical glucocorticoid-treated disease may result from the reduction of CD36 on macrophage membranes. Intriguingly, CD36 is a gustatory lipid sensor 79, whose deficit in cell membranes may account for the symptoms of ageusia, anosmia, diarrhoea, hyperglycaemia, platelet aggregation, and cardiovascular disturbances experienced by Covid-19 patients 75–77. CD36 is a substrate of matrix metalloproteinase-9 (MMP-9) and disintegrin metalloproteinase domain-containing protein 17 (ADAM17) 80, 81. Decreased CD36 expression in Covid-19 patients most likely results from the action of these enzymes. MMP-9 is produced by neutrophils 82, induced by TNF-α 83, unaffected by glucocorticoids, and associated with respiratory syndrome in Covid-19 84. ADAM17 has been described as a target for Covid-19 treatment 85.
In agreement with previous reports 1, 34, 86, 87, in the present study, high concentrations of IL-6, IL-8, and IL-10 were detected, along with insignificant levels of IL-1β in plasma. In contrast, the inflammatory cytokine TNF-α was found at the lowest concentration in BAL fluid, and insignificant levels were detected in the blood. The low levels of TNF-α production are correlated with the decreased counts of intermediate monocytes in the blood and BAL, which are major sources of this cytokine 88. As expected, in the BAL from patients with severe/critical Covid-19, the levels of IL-8, IL-6, IL-1β, TNF-α, and IL-10 were significantly higher compared with the plasma of all participants, or when paired with patients’ own plasma. Notably, in BAL, the most abundant cytokine was IL-8, which is produced by neutrophils 89, lung macrophages 90, and bronchial epithelial cells 10, and is induced by 5-HETE 31 and ACh 10. Interestingly, BAL fluid from non-Covid-19 patients also presented a high concentration of cytokines that did not differ from that of patients with Covid-19. This suggests that cytokine storms are not a key differential factor in this disease. Contrary to our expectations, glucocorticoids did not alter cytokine production (Table S7). As a result, we suggest that, regardless of whether Covid-19 is treated with glucocorticoids, cross-talk occurs between cytokines, lipid mediators, and ACh, as previously reported 2, 3, 91. IL-1β induces 11-HETE 46; arterial relaxation induced by ACh is mediated by 11-HETE and is inhibited by indomethacin 43. In our studies on scorpion envenomation, ACh release was found to be mediated by PGE2-induced by IL-1β and inhibited by indomethacin 3. In the present study, comparing AA, 5-HETE, 11-HETE, cytokine, and receptor expression demonstrated the absence of glucocorticoid effects. These results, in addition to the results of other studies from our laboratory 3, suggest that glucocorticoids did not have an effect in our cohort. This was most likely due to the fact that treatment was started late, namely after inflammation has been triggered by infection. In this context, Covid-19 patients could had already reached the point of no return, similar to what has been previously described in scorpion envenomation 3.
As predicted by Virgilis and Giovani 92, neurotransmitters are produced in Covid-19. ACh induces mucus secretion, bronchoconstriction, lung inflammation and remodelling 9, cardiac dysfunction in scorpionism 3, NETs formation 93, IL-8 release 10, thrombosis 94, and obesity-related severity 95. In addition to the nervous system 4, ACh is also produced by pulmonary vessels 96, airway epithelial cells 6, and immune cells 5. When binding to anti-inflammatory neural nicotinic receptors induces AA release 97, it inhibits nicotine receptors 97, which are highly expressed in lungs 98. In our cohort, a correlation between the AA and ACh levels and the severity of Covid-19 infection was observed, which was reinforced by the results of our bioinformatics study. Our lung transcriptome re-analyses showed that ACh release cycle genes were activated, including acetylcholine-synaptic release (synaptotagmin 1) and neuronal choline transporter (Solute Carrier Family 5 Member 7, SLC5A7). Accordingly, SLC5A7 gene expression was higher in the lung of patients who died from Covid-19 infection than patients who survived 99. Because this gene mediates the translocation of choline into lung epithelial cells 100 and macrophages 101, the production of ACh in BAL may also depend on non-neuronal cells. Indeed, the ACh repressor gene in non-neural cells (RE1 Silencing Transcription Factor) is reduced in the lungs of patients who died of Covid-19 99. The activation of T-lymphocyte EP4 induces the release of ACh 102, suggesting that AA metabolites contribute to the release of ACh from lung and immune cells. As expected, ACh production in patients with severe/critical disease was found to be inhibited by glucocorticoids, since this drug blocks ACh production by lung epithelial cells 103–105. Patients with Covid-19 (mostly methylprednisolone-treated patients) showed lower plasma levels of choline and higher plasma levels of phosphocholine 37. In addition, a distinct pattern of ACh receptor mRNA expression was found in patients who died of Covid-19 (mainly in the CVL-lengthy hospital stay group). This profile is characterised by low levels of expression of the nicotinic receptor encoded by cholinergic receptor nicotinic alpha 7 subunit gene (CHRNA7) and increased levels of expression of two nicotinic and one muscarinic receptor in deceased-CVL compared to deceased non-Covid-19 patients. These proteins are encoded by the cholinergic nicotinic alpha 3 subunit gene (CHRNA3), the cholinergic nicotinic receptor 5 subunit gene (CHRNA5), and the muscarinic cholinergic receptor 3 gene (CHRM3), respectively. These three upregulated genes were also co-expressed and associated with pulmonary inflammatory disorders 9, 106–110. Notably, SARS-CoV-2 spike glycoprotein interacts with the α7 nicotinic acetylcholine receptor, which may compromise the cholinergic anti-inflammatory pathway111.
The ACh-nicotinic receptor, encoded by CHRNA7, is expressed in various lung cells and macrophages 112. It displays extensive anti-inflammatory activities, including the inhibition of pro-inflammatory cytokine production 113, the recruitment of neutrophils 114, and the reduction of CD14 expression in human monocytes 115. The increased degrees of inflammation in the BAL fluid of patients with severe/critical Covid-19 may be associated with the lower levels of anti-inflammatory CHRNA7 expression. However, a reduction in monocyte CD14 may result from high levels of IL-6 production 116, since the levels of CHRNA7 expression were very low. Interestingly, the ACh-M3 receptor in immune cells has been found to be up-regulated by ACh, which mediates its pro-inflammatory actions, including the production of IL-8 and the recruitment of neutrophils117. Furthermore, the interaction between ACh and its receptor triggers the AA-derived release of eicosanoids, including 5-HETE 97, 118, 119. Remarkably, vitamin D modulates the ACh-M3 receptor 120, which may explain its beneficial effects in the prevention of Covid-19 121. Based on the anti-inflammatory effects of nicotinic receptors and the downregulation of the SARS-CoV-2 receptor ACE2 promoted by nicotine, therapies involving nicotinic receptors have been proposed to treat Covid-19 infection 122, 123. In fact, based on these findings, an acetylcholinesterase inhibitor 124, 125 and a nicotinic receptor agonist have been tested 126, 127. However, recent studies have demonstrated that nicotine and smoking increase ACE2 receptor density 128, 129, and that the potential beneficial effects of nicotine were restricted to a small group of individuals 130. Our findings provide a warning that precautions should be taken when considering the therapeutic use of nicotinic agonists, since patients with severe/critical Covid-19 were found to release high amounts of ACh, in addition to showing increased levels of M3-receptor expression. Due to the antiviral and anti-inflammatory effects of AA, it has been proposed to treat Covid-19 patients 60. Nevertheless, AA may be not be an adequate therapeutic target, based on both our findings and previous studies 131, which have shown that AA can favour hyperinflammation and lethality in Covid-19 infection. In conclusion, ACh, AA, 5-HETE, and 11-HETE mediate the innate immune response to SARS-CoV-2 and may define the outcome of infection.
Covid-19 severity can be associated with disruption of homeostatic lipidome and metabolic alterations and this phenomenon may be influenced by comorbidities/risk factors related to the infection. Our findings demonstrated that (i) high plasma levels of ACh and lipid mediators positively correlated with Covid-19 severity; and (ii) only hypertension and/or age were confounding variables for analyzing the association of high plasma levels of AA, 5-HETE, and ACh with disease severity (Table S7). Similarly, the correlation between altered plasma profile of lipid mediators and Covid-19 severity is associated with selective comorbidities – mainly high body mass index (BMI) – but poorly associated with gender, advanced age, and diabetes. However, the correlation between altered AA and/or 5-HETE levels and disease severity is associated with male gender, hypertension, and heart disease, but not BMI53.
We also examined whether glucocorticoid-therapy interfered with the potential effect of some confounding variables on the correlation between high ACh levels and Covid-19 severity in severe/critical patients. Despite the relatively underrepresented samples from non-glucocorticoid treated patients, no confounding variable significantly affected the correlation between plasma and BAL ACh levels and Covid-19 severity in severe/critical patients treated or not with glucocorticoids (Table S8). The altered plasma and BAL levels of the lipid mediators AA, 5-HETE, and 11-HETE in severe/critical patients were associated with the disease severity but not with the confounding variables tested (Table S9). Altogether, the findings here reported suggest that age and/or hypertension are significant confounding variables for analyzing the association between increased levels of cholinergic and lipid mediators and Covid-19 severity. Also, the confounding variables tested probably did not modify the inhibitory action of glucocorticoids on plasma and BAL ACh levels in severe/critical patients.
To the best of our knowledge, this study is the first to demonstrate that the lung inflammatory process and poor outcomes of patients with Covid-19 infection are associated with high levels of lipid mediators and ACh, produced via a partially glucocorticoid-insensitive pathway. Glucocorticoid therapy was found to lower only the levels of ACh. Thus, to improve the benefits of glucocorticoid therapy, we suggest that treatment in hospitalised patients be started early and be preferentially administered to patients with low or undetectable viral loads and harmful lung inflammation in combination with AA metabolism inhibitors.
Conclusions
We demonstrate for the first time that the lung inflammatory process and worse outcomes in Covid-19 are associated with lipid mediators and ACh, which are produced through a partially glucocorticoid-insensitive pathway. To improve the benefits of glucocorticoid therapy, we suggest that it should be started early in severe/critical hospitalized patients and in combination with AA metabolism inhibitors.
Data Availability
All data can be requested by email: faccioli@fcfrp.usp.br
Additional Information
Extended data available in Supplementary Appendixes:
Supplementary Appendix I - Pathways and list of corresponding genes related to acetylcholine and arachidonic acid observed in Reactome pathways and Covid-19 biomarkers.
Supplementary Appendix II - Report of CEMITool results for the gene co- expression modular analysis of lung samples from biopsies of Covid-19 and non-Covid-19 patients.
Supplementary Appendix III - Differential gene expression analysis results between Covid-19-death groups (CV, NCV, CVL and CVH) and Covid-19 samples from patients who underwent glucocorticoid treatment (GC) and not (non-GC).
Supplementary Appendix IV - Betweenness and degree values of genes from biological network constructed based on the principal co-expression module M1.
Funding
Fundação de Amparo a Pesquisa do Estado de São Paulo (FAPESP): #2020/05207-6, #2014/07125-6 and #2015/00658-1 for L.H.F.; #2020/08534-8 for M.M.P.; #2018/22667-0 for C.O.S.S.; #2020/05270-0 for V.D.B. Additional support was provided by the National Council for Scientific and Technological Development (CNPq), the Coordination for the Improvement of Higher Educational Personnel (CAPES-Finance Code 001)), and Pró-Reitora de Pesquisa da Universidade de São Paulo, grant-USP-VIDA.
Conflict of Interest Statement
The authors declare that this research was performed without conflicts of interest or commercial or financial gains.
Supplementary Tables
Supplementary Figures
Acknowledgements
The authors thankfully acknowledge Innovation and Technology Park (Supera), the healthy-participants joining as controls and the positive COVID-19 patients as well as their families. We grieve for all patients who lost their lives as a result of this pandemic, including those who provided us with samples to be able to answer scientific questions and contribute to humanity’s eradication of this disease. We also thank Fabiana Rossetto de Moraes, B.Sc., for the cytometry analysis, Caroline Fontanari, M.Sc., for laboratory and technical support, the ICU team, and all hospital professionals, especially the technicians, nurses, physiotherapists, and biomedical personnel, who collaborated on this work through Hospital Santa Casa de Misericórdia in Ribeirão Preto and Hospital São Paulo in Ribeirão Preto. We are grateful for the indispensable contribution of the Ribeirão Preto Municipal Health Department and the employees of the Serviço de Análises Clínicas (SAC) of the Faculdade de Ciências Farmacêuticas de Ribeirão Preto, USP. We also thank Professors Victor Hugo Aquino Quintana, Ph.D., Márcia Regina von Zeska Kress, Ph.D., and Marcia Eliana da Silva Ferreira, Ph.D. for sharing the BSL2 viral laboratory.
References
- 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.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
- 92.↵
- 93.↵
- 94.↵
- 95.↵
- 96.↵
- 97.↵
- 98.↵
- 99.↵
- 100.↵
- 101.↵
- 102.↵
- 103.↵
- 104.
- 105.↵
- 106.↵
- 107.
- 108.
- 109.
- 110.↵
- 111.↵
- 112.↵
- 113.↵
- 114.↵
- 115.↵
- 116.↵
- 117.↵
- 118.↵
- 119.↵
- 120.↵
- 121.↵
- 122.↵
- 123.↵
- 124.↵
- 125.↵
- 126.↵
- 127.↵
- 128.↵
- 129.↵
- 130.↵
- 131.↵
- 132.↵
- 133.
- 134.
- 135.
- 136.
- 137.↵
- 138.
- 139.
- 140.↵
- 141.↵
- 142.
- 143.↵
- 144.↵
- 145.↵
- 146.↵
- 147.↵