Diagnostic metagenomics from BAL samples of allogenic HSCT patients with pulmonary complications

Adam Grundhoff1, Heinrich Lellek* 2, Malik Alawi3, Klupp Eva4, Christner Martin4, Daniela Indenbirken1, Nicolaus Kröger2, Holger Rohde4, Nicole Fischer4

Author address: 

1Leibniz Institute for Experimental Virology, Heinrich-Petto-Institut, 2Stem Cell Transplantation, 3Bioinformatics Service Facility, 4Institute of Medical Microbiology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany

Abstract: 

Introduction: Despite of improved prophylaxis and conventional microbiological diagnostics the risk for pulmonary complications in of hematopoietic stem cell transplant (HSCT) patients continues to be high. While pneumonia remains a significant cause of mortality after HSCT, detection of an infectious agent fails in more than 50% of cases. Unbiased, non-targeted metagenomic RNA sequencing (UMERS) represents a novel method which can detect not only known, but also distantly related or even novel pathogens of viral, bacterial or parasitic origin. As such, UMERS has the potential to significantly improve the detection of infectious agents in clinical as well as health settings.
Materials (or patients) and methods: We have previously established a streamlined NGS platform specifically dedicated to the detection and analysis of human pathogens in diagnostic samples (Fischer, Rohde EID 2014). We here apply unbiased metagenomic RNA sequencing (UMERS) to 24 diagnostic respiratory specimens (BAL) of HSCT patients with CT confirmed pulmonary lesions.
The performance of UMERS analysis was systematically compared to conventional diagnostic tests including bacterial/fungal culture together with multiplex PCR for human viruses known to cause respiratory diseases (InflA, InflB, hPIV1-4, RSV, Rhinoviruses, Enteroviruses, Metapneumovirus, Bocavirus , Adenoviruses and Coronaviruses including HUK1, NL64, 229E and OC43) and bacteria with the potential to cause pneumonia that are difficult to grow in culture (Bordetella pertussis, Bordedella parapertussis, Legionella, Mycoplasma pneumoniae and Chlamydophila pneumoniae).
Results: UMERS detected in 58% (14/24) of the samples pathogens with the potential to cause respiratory disease. In contrast, conventional diagnostics succeeded in only 37,5% (9/24) of the cases to detect a putative pathogen. All pathogens detected by conventional tests were also faithfully identified by UMERS analysis. Additionally, UMERS revealed a viral infection with human Parainfluenza 3 in one patient: two additional samples showed clear signs of  bacterial infection  (S. pneumoniae, Pseudomonas simiae) and three samples contained sequences of fungal organisms  (Aspergillus fumigatus, Fusarium sp.) which can cause fatal pulmonary complications in immunosuppressed patients.
We furthermore report a cluster of severe pneumonia (n=6) in which the index case was tested negative by conventional diagnostic procedures, while UMERS successfully detected hPIV3 viral sequences. Furthermore, the recovery of whole genome sequences from all six cases by UMERS showed low interspecies sequence diversity between these hPIV3 sequences thereby highly suggesting nosocomial infection and spread of the virus between these patients.
Furthermore, we report a case of S. pneumoniae infection unambiguously detected by UMERS but not by bacterial culturing in one patient with suspected post-transplant malignancies as the cause of pulmonary complications. 
Conclusion: These results illustrate the potential of metagenomic approaches for the diagnostic detection of viral, bacterial and/or fungal sequences from clinical samples. UMERS can significantly contribute to both the improvement of diagnostics as well as support of epidemiological analysis of putative nosocomial infections and thus improve the level of recognition and treatment of pulmonary complications in HSCT patients.

2015

abstract No: 

O053

Full conference title: 

Annual Meeting of European Society for Blood and Marrow Transplantation
    • EBMT 41st (2015)