Lung Immune Cells Show Lasting Epigenetic Changes in Long COVID Patients

Scientists in Sweden have uncovered evidence that immune cells deep within the lungs of people living with post COVID-19 condition, commonly known as long COVID, undergo lasting epigenetic rewiring that is not mirrored in the blood. The study, published in Epigenetics Communications, is the first to track DNA methylation changes over time in lung-derived immune cells from patients with persistent symptoms after SARS-CoV-2 infection, and it points to a possible molecular signature of the condition that standard blood tests would miss entirely.

Post COVID-19 condition is characterized by persistent and highly heterogeneous symptoms that most prominently affect the cardiopulmonary system, including breathlessness, fatigue, exercise intolerance, and dysautonomia. Despite years of research, the biological mechanisms driving these symptoms remain poorly understood. Previous work has documented immune dysregulation and epigenetic alterations in peripheral blood following acute infection, but the lung itself, the primary site of infection and immune activation, has remained largely unexplored at the epigenetic level in this patient group.

The research team, led by Frida Nikesjö, Kristofer Hedman, and Maria Lerm of Linköping University, recruited 13 patients who had experienced persistent symptoms affecting daily life for more than 12 weeks after COVID-19 in 2020 and 2021. Participants were referred to the Department of Clinical Physiology at Linköping University Hospital, where they provided blood and induced sputum samples at study inclusion and again one year later. Induced sputum has been validated as a reliable and biologically informative source of pulmonary immune cells for methylome analysis, allowing the researchers to isolate neutrophil-enriched and macrophage-enriched cell fractions from the lung environment.

At each visit, patients also completed validated questionnaires covering fatigue, quality of life, and dyspnea, and underwent objective physiological testing including symptom-limited cardiopulmonary exercise testing, dynamic spirometry, and single-breath carbon monoxide diffusion capacity measurement. To reduce dimensionality in this small cohort, the team combined questionnaire scores and physiological outcomes into a single symptom-physiology composite variable, categorizing each measure from normal to severely abnormal using clinical reference limits.

Genome-wide DNA methylation profiling was performed using the Illumina Infinium Methylation EPIC 850K array, with rigorous preprocessing that removed failed and cross-reactive probes, sex chromosome sites, and probes containing common single nucleotide polymorphisms. Cell type proportions were estimated using the EpiDISH reference-based deconvolution method, and longitudinal differential methylation was assessed within individuals using the limma framework, with patient identity as a blocking factor and estimated cell proportions, sex, age, body mass index, and smoking status included as covariates. Differentially methylated CpG sites required both an absolute mean methylation difference greater than 0.15 and a false discovery rate-adjusted p-value below 0.05.

The most striking finding was a sharp divergence between compartments. No significant longitudinal DNA methylation changes were detected in peripheral blood mononuclear cells, yet pronounced changes emerged in both lung-derived fractions. In the neutrophil-enriched fraction, the researchers identified 446 differentially methylated CpG sites between the two time points, 70 hypermethylated and 376 hypomethylated, mapping to 274 unique genes. Pathway enrichment analysis across 334 KEGG pathways revealed 43 significantly enriched processes, with Wnt signalling and circadian entrainment among the most prominent. Notably, 87 of the affected genes are known to interact with SARS-CoV-2 proteins according to the BioGRID database.

In the macrophage-enriched fraction, 34 differentially methylated CpG sites were identified, evenly split between hyper- and hypomethylation. Although these changes did not map significantly to any KEGG pathways, an overlap analysis revealed a single gene, KIAA0930, that was differentially methylated in both lung cell fractions. This gene has been implicated in hypoxia adaptation and lung cancer development, and protein interaction analyses place it at a network hub connected to the 14-3-3 protein family, part of the Reactome pathway describing how SARS-CoV-2 targets host intracellular and regulatory pathways. BioGRID further identifies a direct interaction between KIAA0930 and ORF3A, a viral accessory protein known to modulate immune responses and cellular stress pathways.

Crucially, the epigenetic changes were not merely molecular noise. Seventy differentially methylated CpG sites, mapping to 54 genes, showed changes in methylation that correlated with changes in the symptom-physiology composite variable over the year of follow-up. Seven participants improved symptomatically, three remained stable, and three worsened, and their methylation trajectories tracked these clinical courses. Protein interaction and Gene Ontology analyses of the correlated genes revealed significant enrichment in biological pathways related to cardiac function, including a module connected to the KEGG pathway for viral myocarditis, a finding that resonates with the cardiopulmonary symptoms, such as exercise intolerance and dysautonomia, that dominate the clinical picture of post COVID-19 condition.

The authors are careful to frame the study as exploratory and hypothesis-generating. The cohort of 13 patients is small, control samples from pre-pandemic healthy subjects were used only for contextual reference rather than direct statistical comparison, and the lung cell isolation approach yielded enriched rather than fully homogeneous fractions. Neutrophil proportions were higher at the first time point, and residual confounding from comorbidities or treatments cannot be excluded. Without longitudinally matched controls, age-related or time-dependent methylation changes unrelated to post COVID-19 condition cannot be fully ruled out, and no causal relationship between the epigenetic alterations and symptoms can be inferred from these data.

Nevertheless, the findings carry significant implications. The tissue-specific nature of the rewiring suggests that peripheral blood, the workhorse of most long COVID biomarker studies, may be an incomplete window into the disease process, and that neutrophil biology in particular may persist in an altered state well beyond acute infection, consistent with reports of immature and activated neutrophils, neutrophil extracellular trap formation, and epigenetic reprogramming of bone marrow progenitors after COVID-19. The enrichment of pathways involving AMPK signalling, circadian entrainment, and dopaminergic synapses also aligns intriguingly with recent clinical observations, including evidence that the AMPK activator metformin reduces the incidence of post COVID-19 condition. Larger longitudinal studies with matched controls and mechanistic validation will be needed to determine whether these pulmonary epigenetic signatures are drivers of persistent illness, scars of prior infection, or downstream echoes of physiological stress, but this first longitudinal map of the lung methylome in long COVID provides a clear and testable starting point.

Subject of Research: Longitudinal DNA methylation changes in lung immune cells of patients with post COVID-19 condition

Article Title: Longitudinal epigenetic rewiring in lung immune cells in patients with post COVID-19 condition

Article References: Nikesjö, F., Smiljanić, J., Sayyab, S., Martínez-Enguita, D., Gustafsson, M., Rosvall, M., Hedman, K., & Lerm, M. (2026). Longitudinal epigenetic rewiring in lung immune cells in patients with post COVID-19 condition. Epigenetics Communications, 6(1), Article 5. https://doi.org/10.1186/s43682-026-00046-6

Image Credits: AI Generated

DOI: 10.1186/s43682-026-00046-6

Keywords: post COVID-19 condition, long COVID, DNA methylation, epigenetics, lung immune cells, neutrophils, macrophages, induced sputum, SARS-CoV-2, cardiopulmonary symptoms, KIAA0930, viral myocarditis