Assinatura Molecular na Diferenciação da Tonsilite Viral e Bacteriana

Abstract

Introduction: Acute tonsillitis, of viral or bacterial origin, represents a diagnostic challenge. Inadequate distinction between etiologies may lead to the overuse of antibiotics and antimicrobial resistance. Strategies that integrate microbial characterization and the local immune response may improve differential diagnosis. Objective: To identify an in situ biomarker panel capable of differentiating bacterial and viral tonsillitis. Methods: This was a cross-sectional study involving 197 patients seen at Hospital Santa Izabel (Salvador, BA) with clinical suspicion of acute tonsillitis, along with 5 healthy controls. Inclusion into groups was based on the modified Centor Score (bacterial) and the presence of upper respiratory symptoms with tonsillar signs (viral). Clinical classification was performed by emergency physicians and specialists. Patients with recent use of antibiotics, anti-inflammatory drugs, or immunosuppressants were excluded. Following collection of tonsillar swab samples, DNA and RNA were converted into cDNA, prepared using targeted panels, and sequenced on the Illumina MiSeq platform. Analyses included host filtering, quality control, taxonomic annotation, and normalization. Metagenomic analysis was conducted on the CZID platform, with abundance filtering (≥100 rPM) and exclusion of irrelevant species. Resistance genes were identified using ARIBA and ResFinder. Gene expression was evaluated using nCounter technology (NanoString). Transcriptomic analysis included differential expression (log₂), FDR correction, and functional enrichment (GO, KEGG, Reactome) via gProfiler. Data was filtered by prevalence (≥20%) and analyzed using non-parametric tests, PLS-DA, clustering, microorganism-gene correlations, and ROC/PR curves with 3-fold cross-validation. Alpha and beta diversities were evaluated using ecological metrics and PCoA. Results: A total of 46 samples were analyzed (31 viral, 10 bacterial – emergency physicians; 29 viral and 12 bacterial – specialists), along with 5 controls. Metagenomic analysis revealed a complex microbial community in both patients and healthy controls, including pathogens and commensals. There were no significant differences in alpha or beta diversity between groups. Resistance genes were detected regardless of clinical suspicion. Gene expression analysis (n=24) distinguished between acute tonsillitis groups. In bacterial cases, genes such as IL21R, CD47, ITGAM, and CD163 were upregulated, with enrichment of the MHC II pathway. In viral cases, genes such as HLA-DRB5 and TUBA3E were upregulated, with enrichment of the connexon trafficking pathway (p < 0.01). Conclusion: In situ gene expression demonstrated greater diagnostic accuracy than mere pathogen detection, showing promise in etiological differentiation. Metagenomics was limited by the inherent complexity of the analysis and the diversity of the oral cavity microbiota.