Molecular signature of postmortem lung tissue from COVID-19 patients suggests distinct trajectories driving mortality

  1. Anshul Budhraja
  2. Anubhav Basu
  3. Atish Gheware
  4. Dasari Abhilash
  5. Seesandra Rajagopala
  6. Suman Pakala
  7. Madhuresh Sumit
  8. Animesh Ray
  9. Arulselvi Subramaniam
  10. Purva Mathur
  11. Aruna Nambirajan
  12. Sachin Kumar
  13. Ritu Gupta
  14. Naveet Wig
  15. Anjan Trikha
  16. Randeep Guleria
  17. Chitra Sarkar
  18. Ishaan Gupta
  19. Deepali Jain

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

To elucidate the molecular mechanisms that manifest lung abnormalities during severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infections, we performed whole-transcriptome sequencing of lung autopsies from 31 patients with severe COVID-19 and ten uninfected controls. Using metatranscriptomics, we identified the existence of two distinct molecular signatures of lethal COVID-19. The dominant ‘classical’ signature (n=23) showed upregulation of the unfolded protein response, steroid biosynthesis and complement activation, supported by massive metabolic reprogramming leading to characteristic lung damage. The rarer signature (n=8) that potentially represents ‘cytokine release syndrome’ (CRS) showed upregulation of cytokines such as IL1 and CCL19, but absence of complement activation. We found that a majority of patients cleared SARS-CoV-2 infection, but they suffered from acute dysbiosis with characteristic enrichment of opportunistic pathogens such as Staphylococcus cohnii in ‘classical’ patients and Pasteurella multocida in CRS patients. Our results suggest two distinct models of lung pathology in severe COVID-19 patients, which can be identified through complement activation, presence of specific cytokines and characteristic microbiome. These findings can be used to design personalized therapy using in silico identified drug molecules or in mitigating specific secondary infections.

Article activity feed

  1. Version published to 10.1242/dmm.049572
  2. Version published to 10.1101/2021.11.08.467705v2 on bioRxiv
  3. ScreenIT

    SciScore for 10.1101/2021.11.08.467705: (What is this?)

    Please note, not all rigor criteria are appropriate for all manuscripts.

    Table 1: Rigor

    NIH rigor criteria are not applicable to paper type.

    Table 2: Resources

    Software and Algorithms
    SentencesResources
    The NGS library was prepared after cytoplasmic and mitochondrial rRNA depletion, using TruSeq Stranded Total RNA Gold kit per manufacturers’ instructions (Illumina, 20020598).
    NGS
    suggested: (PM4NGS, RRID:SCR_019164)
    Host transcriptome analysis: Raw Illumina sequencing reads were checked for quality using FastQC (version 0.11.9) (Babraham Bioinformatics - FastQC A Quality Control Tool for High Throughput Sequence Data, n.d.) followed by adapter clipping and trimming using Trimmomatic (version 0.39) (Bolger et al., 2014) with default parameters.
    FastQC
    suggested: (FastQC, RRID:SCR_014583)
    Trimmomatic
    suggested: (Trimmomatic, RRID:SCR_011848)
    Trimmed reads were then aligned to the human reference genome (GRCh38, GENCODE v36) (Frankish et al., 2019; Schneider et al., 2017) using STAR aligner (version 2.7.8a) (Dobin et al., 2013).
    GENCODE
    suggested: (GENCODE, RRID:SCR_014966)
    STAR
    suggested: (STAR, RRID:SCR_004463)
    FeatureCounts (subread package version 2.0.1) (Y.
    FeatureCounts
    suggested: (featureCounts, RRID:SCR_012919)
    Quality checks were performed at each step using the MultiQC tool (version 1.10.1) (Ewels et al., 2016).
    MultiQC
    suggested: (MultiQC, RRID:SCR_014982)
    ClusterProfiler package (version 3.18.0) (Yu et al., 2012) was used for the Gene Ontology (GO) term Over Representation Analysis (ORA) of differentially expressed genes.
    ClusterProfiler
    suggested: (clusterProfiler, RRID:SCR_016884)
    GSVA package (version 1.38.2) (Hänzelmann et al., 2013) was used for all GSVA analysis and heatmaps of GSVA enrichment scores were visualized using the package pheatmap (version 1.0.12) (Pheatmap: Pretty Heatmaps, n.d.).
    pheatmap
    suggested: (pheatmap, RRID:SCR_016418)
    Boxplots and other visualizations were made using the ggplot2 package (version 3.3.3) (Wickham, 2011).
    ggplot2
    suggested: (ggplot2, RRID:SCR_014601)
    For identification of transcription factors driving gene expression, we used the Enrichr tool (Kuleshov et al., 2016) using lists of genes upregulated in severe COVID-19 samples (and sub-groups as identified in the paper) when compared with controls.
    Enrichr
    suggested: (Enrichr, RRID:SCR_001575)
    KEGG pathways were utilized from MSigDB (Liberzon et al., 2011).
    KEGG
    suggested: (KEGG, RRID:SCR_012773)
    Connectivity Map (CMap) analysis: Connectivity Map (CMap) (Lamb et al., 2006) analysis was performed using the online portal https://clue.io/cmap to determine perturbagens (potential drugs reversing the aberrant gene expression) using the L1 version of CMap with L1000 data repository, Touchstone data set as a benchmark for assessing connectivity among perturbagens and Individual query option.
    CMap
    suggested: (CMAP, RRID:SCR_009034)
    The filtered unmapped reads were then input into Seal (from the suite of bbtools) and binned into bacterial rRNA (using SILVA bacterial rRNA database) (Quast et al., 2013), human genome (GRCh38) and microbial bin.
    SILVA
    suggested: (SILVA, RRID:SCR_006423)
    The alpha diversity (Shannon diversity index) and bacterial taxon abundance was assessed using the PhyloSeq package (version 1.34.0) (McMurdie & Holmes, 2013).
    PhyloSeq
    suggested: (phyloseq, RRID:SCR_013080)
    Filtered microbial reads from these samples were aligned against the SARS-CoV-2 reference genome (F. Wu et al., 2020) using BBMap (version 38.9) (Bushnell, 2014).
    BBMap
    suggested: (BBmap, RRID:SCR_016965)
    Depth and coverage of the viral genome were obtained using samtools (version 1.9) (Li et al., 2009).
    samtools
    suggested: (SAMTOOLS, RRID:SCR_002105)
    Full length genomes were assembled for samples with high depth and coverage using SPAdes.
    SPAdes
    suggested: (SPAdes, RRID:SCR_000131)
    Transcriptome analysis was performed by aligning filtered viral reads to the reference strain (Wuhan-Hu-1) using Bowtie2 (Langmead & Salzberg, 2012).
    Bowtie2
    suggested: (Bowtie 2, RRID:SCR_016368)

    Results from OddPub: Thank you for sharing your data.

    Results from LimitationRecognizer: An explicit section about the limitations of the techniques employed in this study was not found. We encourage authors to address study limitations.

    Results from TrialIdentifier: No clinical trial numbers were referenced.

    Results from Barzooka: We did not find any issues relating to the usage of bar graphs.

    Results from JetFighter: We did not find any issues relating to colormaps.

    Results from rtransparent:
    • No conflict of interest statement was detected. If there are no conflicts, we encourage authors to explicit state so.
    • No funding statement was detected.
    • No protocol registration statement was detected.

    Results from scite Reference Check: We found no unreliable references.

    About SciScore

    SciScore is an automated tool that is designed to assist expert reviewers by finding and presenting formulaic information scattered throughout a paper in a standard, easy to digest format. SciScore checks for the presence and correctness of RRIDs (research resource identifiers), and for rigor criteria such as sex and investigator blinding. For details on the theoretical underpinning of rigor criteria and the tools shown here, including references cited, please follow this link.

  4. Version published to 10.1101/2021.11.08.467705v1 on bioRxiv