Correlation of the peripheral blood DNA methylation profile and immune‐related adverse events in patients with solid cancer treated with immune checkpoint inhibitors
Markus Kleinberger, Angelika Martina Starzer, Erwin Tomasich, Julia Maria Berger, Vincent Sunder‐Plassmann, Lynn Gottmann, Isabella Solano Henao, Martin Korpan, Josef Fuerst, Nikola Vladic, Cihan Ay, Helmuth Haslacher, Julia Furtner, Matthias Preusser, Gerwin Heller, Anna Sophie BerghoffAbstract
Background/Introduction
Immune‐related adverse events (irAEs) remain unpredictable complications of immune checkpoint inhibitor (ICI) therapy. Current biomarkers, such as inflammatory blood parameter lack specificity and clinical robustness. Because DNA methylation reflects stable immune programming and predisposition to autoimmune diseases, blood methylation profiling may offer a tumour‐agnostic approach for irAE prediction.
Material & methods
We analysed baseline leukocyte DNA methylation in a prospectively enrolled real‐world cohort of advanced pan‐cancer patients treated with ICI in a palliative setting. Genome‐wide methylation was profiled using Infinium MethylationEPIC v1.0 BeadChips. Differentially methylated positions (DMPs) were identified by rank‐based comparison and tested for Gene Ontology enrichment. Blood‐cell deconvolution using EpiDISH was applied to exclude compositional bias.
Results
Among 195 patients (median age 65 years; 34.9% female), 32.8% experienced at least one irAE, most frequently involving thyroid, skin, or lung. Despite globally concordant DNA methylation between groups ( ρ = 0.9996), DMP analysis revealed subtle but distinct site‐specific differences. A methylation‐based classifier discriminated patients who developed irAEs with good accuracy (area under the curve [AUC] 0.831, 95% confidence interval [CI] 0.735–0.927), which remained robust in an internal validation subset (AUC 0.874, 95% CI 0.757–0.992). EpiDISH confirmed no significant variation in immune cell fractions between groups. Functional enrichment of top‐ranked DMPs showed significant overrepresentation of neuronal development, cell‐cell signalling, and epithelial morphogenesis pathways, networks linked to immune‐synapse formation, cytokine signalling, and tissue repair responses.
Discussion
Baseline blood DNA methylation captures systemic immune‐developmental states predisposing to irAEs. The enrichment of cell‐cell signalling and epithelial morphogenesis pathways suggests an epigenetically encoded predisposition for dysregulated immune‐tissue communication. Blood‐based methylation profiling thus represents a promising, tumour‐agnostic biomarker approach for individualised irAE risk prediction in cancer immunotherapy.