Is cellular senescence a biological feature of Long COVID? A transcriptomic analysis across comparative post-acute sequelae phenotypes

INTRODUCTION: Cellular senescence, involving cell-cycle arrest and inflammatory factor release, may play a role in Long COVID development. We investigated the role of senescence-associated genes across multiple post-acute sequelae phenotypes.

MATERIALS AND METHODS: Participants from a cohort study were grouped into post-COVID-19 groups: (i) sensory, fatigue/difficulty thinking, and difficulty breathing/exercise intolerance, identified through a machine learning (ML) analysis of symptom data; (ii) cognitive impairment, measured using a screening cognitive assessment tool (BRACE); and (iii) persistent dyspnea, identified using a symptom-scale known to correlate with a six-minute-walk-test. Post-infection whole blood samples underwent transcriptomic analysis with 47,125 genes used as the reference signature for gene set enrichment analysis (GSEA) against the SenMayo cellular-senescence query gene set (n = 125); enrichment were based on normalized enrichment scores (NES).

RESULTS: SenMayo genes were significantly upregulated in the early post-infection period in (a) 56 individuals who subsequently developed Long COVID symptom phenotypes, compared to 104 who did not develop Long COVID symptoms (NES = 6.04, P = 0.001); (b) 32 individuals with long-term cognitive impairment compared to 40 without long-term cognitive impairment measured by BRACE (NES = 2.58, P = 0.032); and (c) 26 participants who had persistent dyspnea compared to 53 without persistent dyspnea measured by the validated dyspnea instrument (NES = 6.29, P = 0.001). ETS2 was upregulated across all impairment-related phenotypes, while other SenMayo genes showed phenotype-specific variability.

CONCLUSIONS: This study suggests a role of senescence-associated genes in the development of Long COVID, including shared and unique transcriptomic patterns across phenotypes.