Whole Blood RNA-Seq Differentiates Hematopoietic Cell Transplant Recipients with Upper Versus Lower Respiratory Tract Rhinovirus Infection

Abstract

Background Human rhinovirus (HRV) is the most common virus detected in the respiratory tract following hematopoietic cell transplantation (HCT). Mortality following virologically proven lower respiratory tract infection (LRTI) may be up to 40%. Specific immune processes activated in upper respiratory tract infection (URTI) vs LRTI are not well understood. Methods Whole blood was collected in PAXgene tubes at the time of URTI or proven (HRV detected in the lower tract with radiographic changes) or possible (HRV detected in the upper tract with radiographic changes) LRTI. RNA was extracted and globin-reduced (PAXgene RNA kit; GlobinClear). Libraries were prepared with the TruSeq RNA Access kit (Illumina), and were pooled and sequenced (2 × 50bp) on an Illumina HiSeq 2500. Low quality reads were filtered prior to alignment (TopHat), and associated with genes (HTSeq). Normalization and significance testing were done using the Bioconductor package, edgeR. Differential expression was defined as |log2 (ratio)| ≥ 1 (± 2-fold) with the FDR set to 5%. Gene ontology (GO) enrichment analysis of a concise set of biological process terms (DAVID GO Direct) was performed using the Bioconductor package, goseq. Results Seventeen subjects with HRV infection following HCT were included (URTI, N = 7; proven LRTI without copathogens, N = 4; possible LRTI, N = 6). RNA was of sufficient quantity and quality to perform RNA-seq. We demonstrated differential gene expression between HRV URTI and proven LRTI (Figure 1A). Differential gene expression between HRV URTI and possible LRTI was not as robust (Figure 1B). Principle component analysis of the top 500 most variable genes demonstrated separation between subjects with HRV URTI and HRV proven LRTI (Figure 2). GO enrichment analyses demonstrated enrichment of terms associated with adaptive immune responses in the URTI group (Table). Conclusions HRV URTI and proven LRTI have highly differential gene expression patterns, whereas possible LRTI gene expression appears to more closely resemble that of URTI. Gene enrichment analysis demonstrates potential increased immune activation in the URTI subjects, which may lead to better control of infection. Larger studies are needed to verify these findings. Whole blood RNA-seq is feasible and may be a useful tool for demonstrating unique transcriptomic signatures for clinical phenotypes in HCT recipients and for identifying specific pathways in pathogenesis.