HHV-6B detection and host gene expression implicate HHV-6B as pulmonary pathogen after hematopoietic cell transplant

Abstract

Limited understanding of the immunopathogenesis of human herpesvirus 6B (HHV-6B) has prevented its acceptance as a pulmonary pathogen after hematopoietic cell transplant (HCT). In this prospective multicenter study of patients undergoing bronchoalveolar lavage (BAL) for pneumonia after allogeneic HCT, we test blood and BAL fluid (BALF) for HHV-6B DNA and mRNA transcripts associated with lytic infection and perform RNA-seq on paired blood. Among 116 participants, HHV-6B DNA is detected in 37% of BALs, 49% of which also have HHV-6B mRNA detection. We establish HHV-6B DNA viral load thresholds in BALF that are highly predictive of HHV-6B mRNA detection and associated with increased risk for overall mortality and death from respiratory failure. Participants with HHV-6B DNA in BALF exhibit distinct host gene expression signatures, notable for enriched interferon signaling pathways in participants clinically diagnosed with idiopathic pneumonia. These data implicate HHV-6B as a pulmonary pathogen after allogeneic HCT.

Fig. 1. Consort diagram.

Number of enrolled participants, samples, and HHV-6B test results.

Fig. 2. The distribution of BALF HHV-6B DNA and mRNA detection over time and by LRTD category.

Panels (A) and (B) show data from 122 BAL events in 113 participants. The first bar in each time period or category indicates the proportion of BALs with HHV-6B DNA detection; the second bar indicates the proportion with HHV-6B mRNA detection, among those with HHV-6B DNA detection (samples without HHV-6B DNA detection were not tested for mRNA). Panels (C) and (D) show data from 45 BAL events (in 42 participants) that had HHV-6B DNA detection. The boxes represent the interquartile range, the horizontal lines and diamonds within the boxes represent the median and mean, respectively, and the upper and lower whiskers extend to the third and first quartiles plus or minus 1.5 times the interquartile range, respectively. Circles represent data points. LRTD indicates lower respiratory tract disease.

Fig. 3. Kaplan–Meier and cumulative incidence plots of time to overall mortality and death due to respiratory failure, respectively, among patients with and without HHV-6B detection in the BAL at different viral load thresholds.

Participants with HHV-6B DNA ≥ 218 copies/ml (2.3 log₁₀ copies/mL) in BALF based on the threshold predictive of detection of at least one of two mRNA transcripts (n = 27) had increased overall mortality (A) and death from respiratory failure (B). Participants with HHV-6B DNA ≥ 578 copies/ml (2.8 log₁₀ copies/mL) in BALF based on the threshold predictive of detection of both mRNA transcripts (n = 18) had a greater increased risk for overall mortality (C) and death from respiratory failure (D). Deaths due to other causes were considered competing events in the cumulative incidence curves. The first and second log-rank or Gray’s test statistics in the figures depict the unadjusted comparison of the curves at day 30 and day 60 post-BAL, respectively, and were two-sided. Results from adjusted Cox regression models are detailed in Tables S4–S5.

Fig. 4. Genome-wide host expression profiling distinguishes between patients with and without HHV-6B detection.

Panels (A–C) depict multidimensional scaling using principal components analysis (PCA) of the entire transcriptome from whole blood to compare between participants with and without HHV-6B detection, among those with a specific clinically determined LRTD diagnosis. Among 17 participants with a viral pneumonia, PCA analysis suggested segregation between those with HHV-6B detection in the BAL and/or plasma (white highlight) versus those without (A). Among 13 participants with IPS, PCA analysis also suggested segregation between those with HHV-6B detection in the BAL and/or plasma (white highlight) versus those without (B). No clear pattern was seen among 17 participants with fungal pneumonia (C). Panels (D–F) display volcano plots of results from pathway analysis using Gene Set Enrichment Analysis (GSEA) to define and compare transcriptional programs by HHV-6B detection status in the same subgroups as panels (A–C). Each plot depicts gene sets that are either significantly upregulated (to the right of the plot) or downregulated (to the left of the plot) in participants with HHV-6B detection in BALF and/or plasma compared to those without. A false discovery rate (FDR) < 0.05 was used to designate significant enrichment of gene sets. Note: Only one patient in the bacterial subgroup had HHV-6B detection in the BAL or plasma, so this group was not further analyzed.