Identifying gene expression signatures of Mycobacterium abscessus clearance utilising long-read sequencing for both cystic fibrosis and non-cystic fibrosis patients

Abstract

Background: Mycobacteria in the Mycobacterium abscessus group (MABS) are a species of environmental non-tuberculousmycobacteria (NTM) found in soil and water habitats that exhibit high levels of intrinsic multi-drug resistance and can cause both asymptomatic and symptomatic disease in humans [1]. At the same time, NTM are found in approximately 10% of cystic fibrosis (CF) patients [2]. In addition to serology biomarkers, investigation of signatures using long-read sequencing may provide an important adjunct for diagnosing and monitoring MABS pulmonary disease (MABS-PD) infection, disease progression and response to treatment, especially for CF patients. Methods: 67 participants who met eligibility criteria for enrolment at five South East Queensland trial sites were recruited into either the nontreatment Observational cohort (N = 30) or a randomized treatment Interventional cohort (N = 37) as part of the FORMAT platform clinical trial. Blood samples were collected from each patient at Screening (Day 0), Week 6, Week 12 and at Final Outcome (either Week 52 or 62, depending on cohort). Total RNA from 55 whole blood samples, including 15 from CF patients and 39 from Non-CF patients, was extracted and sequenced using the Oxford Nanopore Technologies (ONT) PCR-cDNA Barcoding Kit (SQKPCB114.24) using ONT PromethION technology.10Mreads per samplewere acquired. Differential expression, differential polyadenylation and transcript usage analyses between timepoints, for cohorts of CF and Non-CF patients, were executed using DESeq2, Dorado, and linear mixed-effects model. Additionally, Gene Set Enrichment Analysis (GSEA) was carried out to tell pathway boosting using ClusterProfiler. Results: ONT PromethION provided high quality sequences which achieved 11.12M median of number reads among 55 samples, along with a 99% median mapping rate to the human reference genome. Differential Expression analysis from all timepoints shows 3 genes expressed significantly higher in Non-CF patients compared to CF patients. We also found 7 genes significantly upregulated and 2 downregulated in the Intervention cohort compared to the Observation cohort. GSEA analysis shows pathways such as antigen blindingwere boosted by the intervention (Figure 1). Conclusions: Using high quality long-read sequencing, we discover genes expressed differently between Observation and Intervention groups, and also find several pathways enriched by Intervention treatment. Our study highlights the advantages in utilising long-read ONT sequencing for providing new avenues for detecting transcriptional biomarkers for MABS clearance. High quality evidence obtained from the trial will guide decisions for starting treatment and measuring disease severity in patients with MABS-PD, influencing global practice. Acknowledgements: This funding was supported by a grant from the Australian Medical Research Future Fund (MRF1152249). This abstract is presented on behalf of the FORMAT team. We would like to acknowledge participants in the FORMAT trial.