Systematic evaluation of transcriptomic disease risk and diagnostic biomarker overlap between COVID-19 and tuberculosis: a patient-level meta-analysis
Background
The novel coronavirus, SARS-CoV-2, has increased the burden on healthcare systems already strained by a high incidence of tuberculosis (TB) as co-infection and dual presentation are occurring in syndemic settings. We aimed to understand the interaction between these diseases by profiling COVID-19 gene expression signatures on RNA-sequencing data from TB-infected individuals.
Methods
We performed a systematic review and patient-level meta-analysis by querying PubMed and pre-print servers to derive eligible COVID-19 gene expression signatures from human whole blood (WB), PBMCs or BALF studies. A WB influenza dataset served as a control respiratory disease signature. Three large TB RNA-seq datasets, comprising multiple cohorts from the UK and Africa and consisting of TB patients across the disease spectrum, were chosen to profile these signatures. Putative “COVID-19 risk scores” were generated for each sample in the TB datasets using the TBSignatureProfiler package. Risk was stratified by time to TB diagnosis in progressors and contacts of pulmonary and extra-pulmonary TB. An integrative analysis between TB and COVID-19 single-cell RNA-seq data was performed and a population-level meta-analysis was conducted to identify shared gene ontologies between the diseases and their relative enrichment in COVID-19 disease severity states.
Results
35 COVID-19 gene signatures from nine eligible studies comprising 98 samples were profiled on TB RNA-seq data from 1181 samples from 853 individuals. 25 signatures had significantly higher COVID-19 risk in active TB (ATB) compared with latent TB infection (p <0·005), 13 of which were validated in two independent datasets. FCN1- and SPP1-expressing macrophages enriched in BALF during severe COVID-19 were identified in circulation during ATB. Shared perturbed ontologies included antigen presentation, epigenetic regulation, platelet activation, and ROS/RNS production were enriched with increasing COVID-19 severity. Finally, we demonstrate that the overlapping transcriptional responses may complicate development of blood-based diagnostic signatures of co-infection.
Interpretation
Our results identify shared dysregulation of immune responses in COVID-19 and TB as a dual risk posed by co-infection to COVID-19 severity and TB disease progression. These individuals should be followed up for TB in the months subsequent to SARS-CoV-2 diagnosis.

Dylan Sheerin1, Abhimanyu2, W Evan Johnson3, Anna Coussens1,2,4

1 Infectious Diseases and Immune Defence Division, The Walter & Eliza Hall Institute of Medical Research, Parkville 3279, VIC, Australia.
2 Wellcome Centre for Infectious Diseases in Africa, Institute of Infectious Disease and Molecular Medicine, University of Cape Town, Anzio Rd, Observatory, 7925, Western Cape, South Africa.
3 Division of Computational Biomedicine and Bioinformatics Program, Boston University, Boston, MA, USA; Department of Biostatistics, Boston University, Boston, MA, USA.
4 Department of Medical Biology, University of Melbourne, Parkville 3010, VIC, Australia.

Dylan Sheerin PhD

The Walter & Eliza Hall Institute of Medical Research

Appearances