Epigenetic Fingerprint of the SARS-CoV-2 Infection in the Lung of Lethal COVID-19

A common trait amongst severely affected patients with COVID-19 is lung damage and respiratory failure, the most common cause of death.1Milross L. Majo J. Cooper N. et al.Post-mortem lung tissue: the fossil record of the pathophysiology and immunopathology of severe COVID-19.Lancet Respir Med. 2022; 10: 95-106Abstract Full Text Full Text PDF PubMed Scopus (31) Google Scholar,2Bridges J.P. Vladar E.K. Huang H. et al.Respiratory epithelial cell responses to SARS-CoV-2 in COVID-19.Thorax. 2022; 77: 203-209Crossref PubMed Scopus (83) Google Scholar In addition, the persistence of lung lesions in survivors of COVID-19 could be related to prolonged symptoms.2Bridges J.P. Vladar E.K. Huang H. et al.Respiratory epithelial cell responses to SARS-CoV-2 in COVID-19.Thorax. 2022; 77: 203-209Crossref PubMed Scopus (83) Google Scholar Lung disease begins with an initial stage of early epithelial lesions, edema, and endothelial inflammation, followed by diffuse alveolar damage, before the onset of proliferation and fibrosis.1Milross L. Majo J. Cooper N. et al.Post-mortem lung tissue: the fossil record of the pathophysiology and immunopathology of severe COVID-19.Lancet Respir Med. 2022; 10: 95-106Abstract Full Text Full Text PDF PubMed Scopus (31) Google Scholar,2Bridges J.P. Vladar E.K. Huang H. et al.Respiratory epithelial cell responses to SARS-CoV-2 in COVID-19.Thorax. 2022; 77: 203-209Crossref PubMed Scopus (83) Google Scholar Because epigenetic shifts underpin many human diseases,3Berdasco M. Esteller M. Clinical epigenetics: seizing opportunities for translation.Nat Rev Genet. 2019; 20: 109-127Crossref PubMed Scopus (320) Google Scholar these chemical modifications could be linked to COVID-19 lung pathology. Our data suggest that DNA methylation signatures in the blood of patients with COVID-19 provides clinical value.4Castro de Moura M. Davalos V. Planas-Serra L. et al.Epigenome-wide association study of COVID-19 severity with respiratory failure.EBioMedicine. 2021; 66103339Abstract Full Text Full Text PDF PubMed Scopus (83) Google Scholar,5Davalos V. García-Prieto C.A. Ferrer G. et al.Epigenetic profiling linked to multisystem inflammatory syndrome in children (MIS-C): A multicenter, retrospective study.EclinicalMedicine. 2022; 50101515Abstract Full Text Full Text PDF PubMed Scopus (6) Google Scholar Thus, we sought to characterize the DNA methylation landscape of lung tissues from patients with COVID-19 to unveil targets of epigenetic dysregulation that can explain the described histopathological findings. Clinical data and autopsy samples of lung tissues were retrospectively collected from 36 patients with polymerase chain reaction test-confirmed SARS-CoV-2 infection and lung involvement (COVID-19 group) from April 18, 2020, to April 23, 2021, and 18 individuals without SARS-CoV-2 infection from July 7, 2018, to December 14, 2020 (control group). DNA methylation profiles were established using the Infinium MethylationEPIC Microarray.4Castro de Moura M. Davalos V. Planas-Serra L. et al.Epigenome-wide association study of COVID-19 severity with respiratory failure.EBioMedicine. 2021; 66103339Abstract Full Text Full Text PDF PubMed Scopus (83) Google Scholar,5Davalos V. García-Prieto C.A. Ferrer G. et al.Epigenetic profiling linked to multisystem inflammatory syndrome in children (MIS-C): A multicenter, retrospective study.EclinicalMedicine. 2022; 50101515Abstract Full Text Full Text PDF PubMed Scopus (6) Google Scholar DNA methylation data are available on Gene Expression Omnibus (GEO).6National Library of MedicineGene Expression Omnibus (GEO) repository.https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE224464Date accessed: November 2, 2023Google Scholar Differentially methylated 5'-cytosine-phosphate-guanine-3' (CpG) sites were obtained deriving a linear model with "limma" package including hospital, EPIC sample plate, sentrix ID, and position as covariates. CpG sites with mean beta differences > |0.15| and adjusted P value < .05 were selected. Hierarchical clustering was performed using Euclidean distance and ward.D clustering method. The differentially methylated CpGs were used to establish an epigenetic signature and further validated with external cohorts from the GEO repository (GSE213478, GSE234014, and GSE92511). RNA expression data were obtained using available cell lines.7Iorio F. Knijnenburg T.A. Vis D.J. et al.A landscape of pharmacogenomic interactions in cancer.Cell. 2016; 166: 740-754Abstract Full Text Full Text PDF PubMed Scopus (1196) Google Scholar Tissue deconvolution using a human cell-type methylation atlas8Moss J. Magenheim J. Neiman D. et al.Comprehensive human cell-type methylation atlas reveals origins of circulating cell-free DNA in health and disease.Nat Commun. 2018; 9: 5068Crossref PubMed Scopus (492) Google Scholar was performed to determine the cell type composition of the samples. Immunohistochemistry was developed using anti-NLRP3 polyclonal antibody (MA5-32255, ThermoFisher Scientific). List of candidate CpG sites and genes are deposited at Mendeley Data.9Noguera-Castells A, Parra J, Davalos V. Epigenetic fingerprint of the SARS-CoV-2 infection in the lung of lethal COVID-19. 2023. Mendeley Data, V1, doi: 10.17632/yz6zwm5tk9.1. Accessed November 5, 2023. https://data.mendeley.com/datasets/yz6zwm5tk9/1Google Scholar Lung autopsy specimens from 36 patients diagnosed with COVID-19 were obtained, with clinicopathological features described in Table 1. We also obtained lung samples from 18 individuals without COVID-19 (Table 1). No differences between COVID-19 and control groups were observed for any of the clinicopathological assessed parameters, except those related to cause of death (Table 1). The objective of the analysis was to identify genomic loci with distinct DNA methylation status in the lungs of patients with COVID-19 vs the control cohort (Fig 1A). We unveiled 2,205 CpG sites with a differential methylation status between the COVID-19 and control groups (Methods, Mendeley data9Noguera-Castells A, Parra J, Davalos V. Epigenetic fingerprint of the SARS-CoV-2 infection in the lung of lethal COVID-19. 2023. Mendeley Data, V1, doi: 10.17632/yz6zwm5tk9.1. Accessed November 5, 2023. https://data.mendeley.com/datasets/yz6zwm5tk9/1Google Scholar). Most of the differentially methylated sites were in gene body regions (40.68%), whereas intergenic sites and 5'-end regulatory regions accounted for 33.42% and 24.08%, respectively. According to CpG context, most of the differentially methylated sites (73.74%) were in genomic regions with low density of CpG dinucleotides ("Open Sea"), whereas 26.26% were placed in CpG islands or flanking sequences (CpG shores and shelves). Among these differential CpG sites, 37.28% (n = 822 CpGs) were hypermethylation events in the COVID-19 vs control group, whereas 62.72% (n = 1,383 CpGs) were hypomethylation changes. We selected the 2,205 significantly different DNA methylation sites to provide an epigenomic signature, referred to as EPILUNG. The use of EPILUNG for our 54 lung specimens classified the samples as COVID-19 or non-COVID-19 by hierarchical cluster analysis (Fisher exact test, P < .001) (Fig 1A). Moreover, among these 2,205 CpGs, 23 were distinct between chronic and acute diffuse alveolar damage (DAD) lung pathology within COVID-19 samples (Fisher exact test, P < .001) (Mendeley data9Noguera-Castells A, Parra J, Davalos V. Epigenetic fingerprint of the SARS-CoV-2 infection in the lung of lethal COVID-19. 2023. Mendeley Data, V1, doi: 10.17632/yz6zwm5tk9.1. Accessed November 5, 2023. https://data.mendeley.com/datasets/yz6zwm5tk9/1Google Scholar). We further validated the specificity of the EPILUNG profile by showing how it also discriminated between our lung COVID-19 cases and two independent cohorts of normal lungs from the pre-COVID-19 era (GEO Repository GSE213478 and GSE234014) (Fisher exact test, both P < .001) (Fig 1B). Furthermore, to prove the specificity of the EPILUNG profile to discriminate a general lung-immune response from a specific COVID-19 lung-immune response, we used another external cohort from individuals who do not use tobacco, individuals who do use tobacco, and individuals with COPD (GEO Repository GSE92511). The clustering analysis proved that EPILUNG signature had a high degree of specificity from patients with COVID-19 disease (Fisher exact test, P < .001) (Fig 1B).Table 1Clinicopathological Characteristics of the Studied Patients With COVID-19 and Control GroupCharacteristicsCOVID-19 CohortControl CohortP ValuecP values were calculated using Fisher exact test or Mann-Whitney test for dichotomous or continuous variables, respectively. P values under .05 represent statistical significant association between covariables.(N = 36)aInclusion criteria for the COVID-19 cohort: Patients >18 years old, polymerase chain reaction test positive for SARS-CoV-2 with complete clinical information of disease history, comorbidities, and follow-up, showing clinical pulmonary involvement and COVID-19-related death.(N = 18)bInclusion criteria for the control cohort: > 18-year-old individuals with complete clinical information about comorbidities, without clinical evidence of SARS-CoV-2 infection, and sudden death caused by cardiopathies, except one case died of cancer dissemination and one postsurgery.Sex, Frequency (%) Female9 (25.0)6 (33.3).536 Male27 (75.0)12 (66.7)Age, y, median [range]68.5 [40-91]70.5 [35-87].941Underlying conditions, frequency (%) Smoking9 (36.0)3 (17.6).300 Hypertension16 (44.4)7 (38.9).776 Diabetes mellitus5 (13.9)4 (22.2).461 Obesity11 (34.4)4 (22.2).523 Respiratory diseasedCOPD or asthma.10 (27.8)4 (22.2).751 Cardiac diseaseeCoronary artery disease, heart failure, or atrial fibrillation.17 (47.2)4 (22.2).138 Chronic kidney disease3 (8.3)5 (27.8).100 Chronic neurological or neuromuscular disease7 (19.4)3 (16.7)1.000 Cancer10 (27.8)2 (11.1).298 Immunocompromised statefImmunocompromised state caused by autoimmune disease or other cause.5 (13.9)4 (22.2).461No. of comorbidities, frequency (%) 0-219 (52.8)12 (66.7).392 3-617 (47.2)6 (33.3)Pulmonary disease, frequency (%) Acute DAD14 (38.9)NANA Chronic DAD22 (61.1)NANACause of death, frequency (%) Multiorgan failure/respiratory distress26 (72.2)0 (0)< .001 Septic shock4 (11.1)0 (0).289 Intestinal necrosis1 (2.8)0 (0)1.000 Pancreatitis2 (5.6)0 (0).547 CardiopathygAcute myocardial infarction, cardiac fibrosis, aortic dissection, hemopericardium, or myocarditis.2 (5.6)13 (72.2)< .001 Hypovolemic shock1 (2.8)0 (0)1.000 Cancer dissemination0 (0)1 (5.6).333 Postsurgery0 (0)1 (5.6).333 Unknown0 (0)3 (16.7)NADAD = diffuse alveolar damage; NA = not applicable.a Inclusion criteria for the COVID-19 cohort: Patients >18 years old, polymerase chain reaction test positive for SARS-CoV-2 with complete clinical information of disease history, comorbidities, and follow-up, showing clinical pulmonary involvement and COVID-19-related death.b Inclusion criteria for the control cohort: > 18-year-old individuals with complete clinical information about comorbidities, without clinical evidence of SARS-CoV-2 infection, and sudden death caused by cardiopathies, except one case died of cancer dissemination and one postsurgery.c P values were calculated using Fisher exact test or Mann-Whitney test for dichotomous or continuous variables, respectively. P values under .05 represent statistical significant association between covariables.d COPD or asthma.e Coronary artery disease, heart failure, or atrial fibrillation.f Immunocompromised state caused by autoimmune disease or other cause.g Acute myocardial infarction, cardiac fibrosis, aortic dissection, hemopericardium, or myocarditis. Open table in a new tab DAD = diffuse alveolar damage; NA = not applicable. Among the 2,205 loci, 1,377 (62.45%) CpG sites were located within 1,085 unique known coding genes, whereas 828 (37.55%) were placed in noncoding sequences. Then, we performed a deconvolution analysis to estimate cell-type fractions in each sample and found that most of the samples were enriched in immune and endothelial cells (38.18% and 18.31%, respectively). Therefore, using a set of 100 blood-derived cell lines characterized for RNA expression and DNA methylation,7Iorio F. Knijnenburg T.A. Vis D.J. et al.A landscape of pharmacogenomic interactions in cancer.Cell. 2016; 166: 740-754Abstract Full Text Full Text PDF PubMed Scopus (1196) Google Scholar upgraded for the EPIC microarray, we found that for 184 unique CpG sites (160 unique genes), the methylation status was significantly associated with transcript levels. Eighty-six CpGs showed a positive correlation with gene expression (r > 0.3, Pearson test, P < .05), whereas 98 CpGs shown a negative correlation (r < −0.3, Pearson test, P < .05) (Mendeley data9Noguera-Castells A, Parra J, Davalos V. Epigenetic fingerprint of the SARS-CoV-2 infection in the lung of lethal COVID-19. 2023. Mendeley Data, V1, doi: 10.17632/yz6zwm5tk9.1. Accessed November 5, 2023. https://data.mendeley.com/datasets/yz6zwm5tk9/1Google Scholar). We selected for additional expression analysis by immunohistochemistry of the NLRP3 gene, a top candidate target (Mendeley data9Noguera-Castells A, Parra J, Davalos V. Epigenetic fingerprint of the SARS-CoV-2 infection in the lung of lethal COVID-19. 2023. Mendeley Data, V1, doi: 10.17632/yz6zwm5tk9.1. Accessed November 5, 2023. https://data.mendeley.com/datasets/yz6zwm5tk9/1Google Scholar). In line with NLRP3 3'-UTR CpG hypermethylation in COVID-19 cases and its association with high levels of the NLRP3 transcript (Mendeley data9Noguera-Castells A, Parra J, Davalos V. Epigenetic fingerprint of the SARS-CoV-2 infection in the lung of lethal COVID-19. 2023. Mendeley Data, V1, doi: 10.17632/yz6zwm5tk9.1. Accessed November 5, 2023. https://data.mendeley.com/datasets/yz6zwm5tk9/1Google Scholar), we found NLRP3 overexpression in COVID-19 patients in comparison with healthy control subjects, particularly within macrophages (Fig 1C). Finally, to unfold the biological functions of the identified 160 genes associated with the COVID-19 lung DNA methylation profile, we data mined gene sets by Gene Ontology, MSigDB Hallmarks, KEGG, and Panther databases using Gene Set Enrichment Analysis. The results showed an enrichment for "chemokine," "interferon gamma," "inflammation," "transforming growth factor-beta," and "Wnt" signaling pathways, also an enrichment in genes whose translated proteins were in the endosomal-lysosomal system (false discovery rate < 0.05) (Mendeley data9Noguera-Castells A, Parra J, Davalos V. Epigenetic fingerprint of the SARS-CoV-2 infection in the lung of lethal COVID-19. 2023. Mendeley Data, V1, doi: 10.17632/yz6zwm5tk9.1. Accessed November 5, 2023. https://data.mendeley.com/datasets/yz6zwm5tk9/1Google Scholar). Thus, the characterized genes and pathways support that lung tissues of patients with COVID-19 undergo a DNA methylation shift associated with disrupted endosomal-lysosomal system homeostasis within a microenvironment of enhanced fibrosis and marked immune response and hyperinflammatory state. Our results dissect for what we believe is the first time the DNA methylation signature that defines the lungs of COVID-19 patients undergoing respiratory failure. Our COVID-19 signature, EPILUNG, is enriched in genes related to inflammation, cell adhesion, fibroblast biology, or the degree of activation of immune cells, which are pathways already described to be altered in COVID-19.1Milross L. Majo J. Cooper N. et al.Post-mortem lung tissue: the fossil record of the pathophysiology and immunopathology of severe COVID-19.Lancet Respir Med. 2022; 10: 95-106Abstract Full Text Full Text PDF PubMed Scopus (31) Google Scholar,2Bridges J.P. Vladar E.K. Huang H. et al.Respiratory epithelial cell responses to SARS-CoV-2 in COVID-19.Thorax. 2022; 77: 203-209Crossref PubMed Scopus (83) Google Scholar,10Melms J.C. Biermann J. Huang H. et al.A molecular single-cell lung atlas of lethal COVID-19.Nature. 2021; 595: 114-119Crossref PubMed Scopus (305) Google Scholar,11Delorey T.M. Ziegler C.G.K. Heimberg G. et al.COVID-19 tissue atlases reveal SARS-CoV-2 pathology and cellular targets.Nature. 2021; 595: 107-113Crossref PubMed Scopus (419) Google Scholar For example, among top candidates genes (Mendeley data9Noguera-Castells A, Parra J, Davalos V. Epigenetic fingerprint of the SARS-CoV-2 infection in the lung of lethal COVID-19. 2023. Mendeley Data, V1, doi: 10.17632/yz6zwm5tk9.1. Accessed November 5, 2023. https://data.mendeley.com/datasets/yz6zwm5tk9/1Google Scholar), we found the zinc finger protein 36 L1 (ZFP36L1) that regulates T-cell responses and is underexpressed in transitioning monocyte-derived macrophages from lung COVID-19 autopsies.10Melms J.C. Biermann J. Huang H. et al.A molecular single-cell lung atlas of lethal COVID-19.Nature. 2021; 595: 114-119Crossref PubMed Scopus (305) Google Scholar Herein, we show hypermethylation status linked to downregulation (Mendeley data9Noguera-Castells A, Parra J, Davalos V. Epigenetic fingerprint of the SARS-CoV-2 infection in the lung of lethal COVID-19. 2023. Mendeley Data, V1, doi: 10.17632/yz6zwm5tk9.1. Accessed November 5, 2023. https://data.mendeley.com/datasets/yz6zwm5tk9/1Google Scholar), providing biological plausibility to our data. Additional genes, such as VPS41, CIITA, TLN1, CTSZ, CD68, LAPTM5, MGLL, and PRKCB (Mendeley data9Noguera-Castells A, Parra J, Davalos V. Epigenetic fingerprint of the SARS-CoV-2 infection in the lung of lethal COVID-19. 2023. Mendeley Data, V1, doi: 10.17632/yz6zwm5tk9.1. Accessed November 5, 2023. https://data.mendeley.com/datasets/yz6zwm5tk9/1Google Scholar) also showed upregulated expression patterns in COVID-19-associated monocytes/macrophages.10Melms J.C. Biermann J. Huang H. et al.A molecular single-cell lung atlas of lethal COVID-19.Nature. 2021; 595: 114-119Crossref PubMed Scopus (305) Google Scholar Our data, in addition to shedding light on the role of epigenetically regulated genes in COVID-19 lung damage, pinpoint specific DNA methylation events that might result in long-term complications. The genes that are differentially methylated mainly are involved in chronic inflammation, vascular imbalance, and fibrosis, which are the primary events leading to the long-term clinical presentation of COVID-19.12George P.M. Wells A.U. Jenkins R.G. Pulmonary fibrosis and COVID-19: the potential role for antifibrotic therapy.Lancet Respir Med. 2020; 8: 807-815Abstract Full Text Full Text PDF PubMed Scopus (753) Google Scholar Therefore, remodeling this DNA methylation landscape by epigenetic agents could potentially be a strategy to mitigate the fibrotic environment seen in long-term COVID-19 cases. Because DNA methylation inhibitors are used in hematologic malignancies, these agents could be considered to treat the severe disorder, as currently assessed on the first clinical trial of this type.13National Institutes of Health Clinical CenterDecitabine for Coronavirus (COVID-19) pneumonia: acute respiratory distress syndrome (ARDS) treatment: DART trial (DART). NCT04482621. ClinicalTrials.gov. National Institutes of Health, 2020http://clinicaltrials.gov/ct2/show/NCT04482621Google Scholar Supported by Fundacio La Marato de TV3 Reference Number 202131-32.