Systematic comparison of published host gene expression signatures for bacterial/viral discrimination

A Four-Gene Signature from Blood to Exclude Bacterial Etiology of Lower Respiratory Tract Infection in Adults

Unnecessary antibiotic use is a major driver of antimicrobial resistance, an urgent public health threat. There is an unmet need for improved diagnostics for identifying bacterial etiology in acute respiratory infection (ARI). Hospitalized adults with ARI underwent comprehensive microbiologic testing and those with definitive viral (n = 280), bacterial (n = 129), or mixed viral-bacterial infection (n = 95) had whole blood RNA sequencing. A hard-thresholded, mostly relaxed, LASSO-constrained logistic regression model was used to select a parsimonious gene set (ITGB4, ITGA7, IFI27, FAM20A) highly capable of discriminating any bacterial from nonbacterial infection (cross validated AUC = 0.90). The 4-gene signature was validated in two independent cohorts (AUC = 0.90-0.94). Thresholding the 4-gene risk score to yield 90% sensitivity to detect bacterial infection resulted in 71% specificity and 91% negative predictive value. This 4-gene signature defining the absence of bacterial ARI may supplement clinical judgement for management of antibiotics in ARI.

Pathogen class-specific transcriptional responses derived from PBMCs accurately discriminate between fungal, bacterial, and viral infections

Immune responses during acute infection often contain canonical elements which are shared across the responses to an array of agents within a given pathogen class (i.e., respiratory viral infection). Identification of these shared, canonical elements across similar infections offers the potential for impacting development of novel diagnostics and therapeutics. In this way, analysis of host gene expression patterns ('signatures') in white blood cells has been shown to be useful for determining the etiology of some acute viral and bacterial infections. In order to study conserved immune elements shared across the host response to related pathogens, we performed in vitro human PBMC challenges with common fungal pathogens (Candida albicans, Cryptococcus neoformans and gattii); four strains of influenza virus (Influenza A/Puerto Rico/08/34 [H1N1, PR8], A/Brisbane/59/2007 [H1N1], A/Solomon Islands/3/2006 [H1N1], and A/Wisconsin/67/2005 [H3N2]); and gram-negative (Escherichia coli) and gram-positive (Streptococcus pneumoniae) bacteria. Exposed human cells were then analyzed for differential gene expression utilizing Affymetrix microarrays. Analysis of pathogen exposure of PBMCs revealed strong, conserved gene expression patterns representing these canonical immune response elements to each broad pathogen class. A 41-gene multinomial signature was developed which correctly classified fungal, viral, or bacterial exposure with 94-98% accuracy. Furthermore, a 21-gene signature consisting of a subset of the discriminatory PBMC-derived genes was capable of accurately differentiating human patients with invasive candidiasis, acute viral infection, or bacterial infection (AUC 0.94, 0.83, and 0.96 respectively). These data reinforce the conserved nature of the genomic responses in human peripheral blood cells upon exposure to infectious agents and highlight the potential for in vitro models to augment our ability to develop novel diagnostic classifiers for acute infectious diseases, particularly devastating fungal infections.

Transfer RNA-derived fragment production in calves challenged with Mycoplasma bovis or co-infected with bovine viral diarrhea virus and Mycoplasma bovis in several tissues and blood

Understanding the molecular mechanisms underlying immune response can allow informed decisions in drug or vaccine development, and aid in the identification of biomarkers to predict exposure or evaluate treatment efficacy. The objective of this study was to identify differentially expressed transfer RNA-derived fragments (tRFs) in calves challenged with Mycoplasma bovis (M. bovis) or co-infected with M. bovis and bovine viral diarrhea virus (BVDV). Serum, white blood cells (WBC), liver, mesenteric lymph node (MLN), tracheal-bronchial lymph node (TBLN), spleen, and thymus were collected from Control (n = 2), M. bovis (MB; n = 3), and co-infected (Dual; n = 3) animals, and small RNAs extracted for sequencing. An average of 94% of reads were derived from 5` halves and/or 5` tRFs in serum, liver, WBC, TBLN, spleen, MLN, and thymus. The expression of tRFs in lymphatic tissues (MLN, TBLN, Thymus, Spleen) were highly correlated with each other (r ≥ 0.82), but not with serum and WBC. A total of 25 and 65 differentially expressed tRFs were observed in liver and thymus, respectively. There were no differentially expressed tRFs found in other tissues analyzed. Nineteen thymus tRFs were differentially expressed in Dual compared to Control and MB, and the predicted targets of these tRFs were associated with MAPK signaling pathways and ERK1 and ERK2 cascades. The differentially expressed tRFs found in thymus and liver may underlie mechanisms of thymic depletion or liver inflammation previously observed in BVDV. Additional studies should be pursued to investigate differential expression of the predicted tRF targets.

Human Conjunctival Transcriptome in Acanthamoeba Keratitis: An Exploratory Study

PURPOSE

The purpose of this study was to identify conjunctival transcriptome differences in patients with Acanthamoeba keratitis compared with keratitis with no known associated pathogen.

METHODS

The host conjunctival transcriptome of 9 patients with Acanthamoeba keratitis (AK) is compared with the host conjunctival transcriptome of 13 patients with pathogen-free keratitis. Culture and/or confocal confirmed Acanthamoeba in 8 of 9 participants with AK who underwent metagenomic RNA sequencing as the likely pathogen. Cultures were negative in all 13 cases where metagenomic RNA sequencing did not identify a pathogen.

RESULTS

Transcriptome analysis identified 36 genes differently expressed between patients with AK and patients with presumed sterile, or pathogen-free, keratitis. Gene enrichment analysis revealed that some of these genes participate in several biologic pathways important for cellular signaling, ion transport and homeostasis, glucose transport, and mitochondrial metabolism. Notable relatively differentially expressed genes with potential relevance to Acanthamoeba infection included CPS1 , SLC35B4 , STEAP2 , ATP2B2 , NMNAT3 , and AKAP12 .

CONCLUSIONS

This research suggests that the local transcriptome in Acanthamoeba keratitis may be sufficiently robust to be detected in the conjunctiva and that corneas infected with Acanthamoeba may be distinguished from the inflamed cornea where no pathogen was identified. Given the low sensitivity for corneal cultures, identification of differentially expressed genes may serve as a suggestive transcriptional signature allowing for a complementary diagnostic technique to identify this blinding parasite. Knowledge of differentially expressed genes may also direct investigation of disease pathophysiology and suggest novel pathways for therapeutic targets.

Integration and validation of host transcript signatures, including a novel 3-transcript tuberculosis signature, to enable one-step multiclass diagnosis of childhood febrile disease

BACKGROUND

Whole blood host transcript signatures show great potential for diagnosis of infectious and inflammatory illness, with most published signatures performing binary classification tasks. Barriers to clinical implementation include validation studies, and development of strategies that enable simultaneous, multiclass diagnosis of febrile illness based on gene expression.

METHODS

We validated five distinct diagnostic signatures for paediatric infectious diseases in parallel using a single NanoString nCounter® experiment. We included a novel 3-transcript signature for childhood tuberculosis, and four published signatures which differentiate bacterial infection, viral infection, or Kawasaki disease from other febrile illnesses. Signature performance was assessed using receiver operating characteristic curve statistics. We also explored conceptual frameworks for multiclass diagnostic signatures, including additional transcripts found to be significantly differentially expressed in previous studies. Relaxed, regularised logistic regression models were used to derive two novel multiclass signatures: a mixed One-vs-All model (MOVA), running multiple binomial models in parallel, and a full-multiclass model. In-sample performance of these models was compared using radar-plots and confusion matrix statistics.

RESULTS

Samples from 91 children were included in the study: 23 bacterial infections (DB), 20 viral infections (DV), 14 Kawasaki disease (KD), 18 tuberculosis disease (TB), and 16 healthy controls. The five signatures tested demonstrated cross-platform performance similar to their primary discovery-validation cohorts. The signatures could differentiate: KD from other diseases with area under ROC curve (AUC) of 0.897 [95% confidence interval: 0.822-0.972]; DB from DV with AUC of 0.825 [0.691-0.959] (signature-1) and 0.867 [0.753-0.982] (signature-2); TB from other diseases with AUC of 0.882 [0.787-0.977] (novel signature); TB from healthy children with AUC of 0.910 [0.808-1.000]. Application of signatures outside of their designed context reduced performance. In-sample error rates for the multiclass models were 13.3% for the MOVA model and 0.0% for the full-multiclass model. The MOVA model misclassified DB cases most frequently (18.7%) and TB cases least (2.7%).

CONCLUSIONS

Our study demonstrates the feasibility of NanoString technology for cross-platform validation of multiple transcriptomic signatures in parallel. This external cohort validated performance of all five signatures, including a novel sparse TB signature. Two exploratory multi-class models showed high potential accuracy across four distinct diagnostic groups.

Predictive signature of murine and human host response to typical and atypical pneumonia

BACKGROUND

Pneumonia due to typical bacterial, atypical bacterial and viral pathogens can be difficult to clinically differentiate. Host response-based diagnostics are emerging as a complementary diagnostic strategy to pathogen detection.

METHODS

We used murine models of typical bacterial, atypical bacterial and viral pneumonia to develop diagnostic signatures and understand the host's response to these types of infections. Mice were intranasally inoculated with Streptococcus pneumoniae, Mycoplasma pneumoniae, influenza or saline as a control. Peripheral blood gene expression analysis was performed at multiple time points. Differentially expressed genes were used to perform gene set enrichment analysis and generate diagnostic signatures. These murine-derived signatures were externally validated in silico using human gene expression data. The response to S. pneumoniae was the most rapid and robust.

RESULTS

Mice infected with M. pneumoniae had a delayed response more similar to influenza-infected animals. Diagnostic signatures for the three types of infection had 0.94-1.00 area under the receiver operator curve (auROC). Validation in five human gene expression datasets revealed auROC of 0.82-0.96.

DISCUSSION

This study identified discrete host responses to typical bacterial, atypical bacterial and viral aetiologies of pneumonia in mice. These signatures validated well in humans, highlighting the conserved nature of the host response to these pathogen classes.

A supervised Bayesian factor model for the identification of multi-omics signatures

MOTIVATION

Predictive biological signatures provide utility as biomarkers for disease diagnosis and prognosis, as well as prediction of responses to vaccination or therapy. These signatures are identified from high-throughput profiling assays through a combination of dimensionality reduction and machine learning techniques. The genes, proteins, metabolites, and other biological analytes that compose signatures also generate hypotheses on the underlying mechanisms driving biological responses, thus improving biological understanding. Dimensionality reduction is a critical step in signature discovery to address the large number of analytes in omics datasets, especially for multi-omics profiling studies with tens of thousands of measurements. Latent factor models, which can account for the structural heterogeneity across diverse assays, effectively integrate multi-omics data and reduce dimensionality to a small number of factors that capture correlations and associations among measurements. These factors provide biologically interpretable features for predictive modeling. However, multi-omics integration and predictive modeling are generally performed independently in sequential steps, leading to suboptimal factor construction. Combining these steps can yield better multi-omics signatures that are more predictive while still being biologically meaningful.

RESULTS

We developed a supervised variational Bayesian factor model that extracts multi-omics signatures from high-throughput profiling datasets that can span multiple data types. Signature-based multiPle-omics intEgration via lAtent factoRs (SPEAR) adaptively determines factor rank, emphasis on factor structure, data relevance and feature sparsity. The method improves the reconstruction of underlying factors in synthetic examples and prediction accuracy of coronavirus disease 2019 severity and breast cancer tumor subtypes.

AVAILABILITY AND IMPLEMENTATION

SPEAR is a publicly available R-package hosted at https://bitbucket.org/kleinstein/SPEAR.

Quantitative and Longitudinal Assessment of Systemic Innate Immunity in Health and Disease Using a 2D Gene Model

Dysregulation of innate immunity is deeply involved in infectious and autoimmune diseases. For a better understanding of pathogenesis and improved management of these diseases, it is of vital importance to implement convenient monitoring of systemic innate immunity. Built upon our previous works on the host transcriptional response to infection in peripheral blood, we proposed a 2D gene model for the simultaneous assessment of two major components of systemic innate immunity, including VirSig as the signature of the host response to viral infection and BacSig as the signature of the host response to bacterial infection. The revelation of dysregulation in innate immunity by this 2D gene model was demonstrated with a wide variety of transcriptome datasets. In acute infection, distinctive patterns of VirSig and BacSig activation were observed in viral and bacterial infection. In comparison, both signatures were restricted to a defined range in the vast majority of healthy adults, regardless of age. In addition, BacSig showed significant elevation during pregnancy and an upward trend during development. In tuberculosis (TB), elevation of BacSig and VirSig was observed in a significant portion of active TB patients, and abnormal BacSig was also associated with a longer treatment course. In cystic fibrosis (CF), abnormal BacSig was observed in a subset of patients, and no overall change in BacSig abnormality was observed after the drug treatment. In systemic sclerosis-associated interstitial lung disease (SSc-ILD), significant elevation of VirSig and BacSig was observed in some patients, and treatment with a drug led to the further deviation of BacSig from the control level. In systemic lupus erythematosus (SLE), positivity for the anti-Ro autoantibody was associated with significant elevation of VirSig in SLE patients, and the additive effect of VirSig/BacSig activation was also observed in SLE patients during pregnancy. Overall, these data demonstrated that the 2D gene model can be used to assess systemic innate immunity in health and disease, with the potential clinical applications including patient stratification, prescription of antibiotics, understanding of pathogenesis, and longitudinal monitoring of treatment response.

Distinguishing community-acquired bacterial and viral meningitis: Microbes and biomarkers

Diagnostic tools to differentiate between community-acquired bacterial and viral meningitis are essential to target the potentially lifesaving antibiotic treatment to those at greatest risk and concurrently spare patients with viral meningitis from the disadvantages of antibiotics. In addition, excluding bacterial meningitis and thus decreasing antibiotic consumption would be important to help reduce antimicrobial resistance and healthcare expenses. The available diagnostic laboratory tests for differentiating bacterial and viral meningitis can be divided microbiological pathogen-focussed methods and biomarkers of the host response. Bacterial culture-independent microbiological methods, such as highly multiplexed nucleic acid amplification tests, are rapidly making their way into the clinical practice. At the same time, more conventional host protein biomarkers, such as procalcitonin and C-reactive protein, are supplemented by newer proteomic and transcriptomic signatures. This review aims to summarise the current state and the recent advances in diagnostic methods to differentiate bacterial from viral meningitis.

RNA sequencing reveals differential long noncoding RNA expression profiles in bacterial and viral meningitis in children

BACKGROUND

We aimed to investigate the involvement of long non-coding RNA (lncRNA) in bacterial and viral meningitis in children.

METHODS

The peripheral blood of five bacterial meningitis patients, five viral meningitis samples, and five healthy individuals were collected for RNA sequencing. Then, the differentially expressed lncRNA and mRNA were detected in bacterial meningitis vs. controls, viral meningitis vs. healthy samples, and bacterial vs. viral meningitis patients. Besides, co-expression and the competing endogenous RNA (ceRNA) networks were constructed. Receiver operating characteristic curve (ROC) analysis was performed.

RESULTS

Compared with the control group, 2 lncRNAs and 32 mRNAs were identified in bacterial meningitis patients, and 115 lncRNAs and 54 mRNAs were detected in viral meningitis. Compared with bacterial meningitis, 165 lncRNAs and 765 mRNAs were identified in viral meningitis. 2 lncRNAs and 31 mRNAs were specific to bacterial meningitis, and 115 lncRNAs and 53 mRNAs were specific to viral meningitis. The function enrichment results indicated that these mRNAs were involved in innate immune response, inflammatory response, and immune system process. A total of 8 and 1401 co-expression relationships were respectively found in bacterial and viral meningitis groups. The ceRNA networks contained 1 lncRNA-mRNA pair and 4 miRNA-mRNA pairs in viral meningitis group. GPR68 and KIF5C, identified in bacterial meningitis co-expression analysis, had an area under the curve (AUC) of 1.00, while the AUC of OR52K2 and CCR5 is 0.883 and 0.698, respectively.

CONCLUSIONS

Our research is the first to profile the lncRNAs in bacterial and viral meningitis in children and may provide new insight into understanding meningitis regulatory mechanisms.

Whole blood stimulation as a tool for studying the human immune system

The human immune system is best accessible via tissues and organs not requiring major surgical intervention, such as blood. In many circumstances, circulating immune cells correlate with an individual's health state and give insight into physiological and pathophysiological processes. Stimulating whole blood ex vivo is a powerful tool to investigate immune responses. In the context of clinical research, the applications of whole blood stimulation include host immunity, disease characterization, diagnosis, treatment, and drug development. Here, we summarize different setups and readouts of whole blood assays and discuss applications for preclinical research and clinical practice. Finally, we propose combining whole blood stimulation with high-throughput technologies, such as single-cell RNA-sequencing, to comprehensively analyze the human immune system for the identification of biomarkers, therapeutic interventions as well as companion diagnostics.

Host-response transcriptional biomarkers accurately discriminate bacterial and viral infections of global relevance

Diagnostic limitations challenge management of clinically indistinguishable acute infectious illness globally. Gene expression classification models show great promise distinguishing causes of fever. We generated transcriptional data for a 294-participant (USA, Sri Lanka) discovery cohort with adjudicated viral or bacterial infections of diverse etiology or non-infectious disease mimics. We then derived and cross-validated gene expression classifiers including: 1) a single model to distinguish bacterial vs. viral (Global Fever-Bacterial/Viral [GF-B/V]) and 2) a two-model system to discriminate bacterial and viral in the context of noninfection (Global Fever-Bacterial/Viral/Non-infectious [GF-B/V/N]). We then translated to a multiplex RT-PCR assay and independent validation involved 101 participants (USA, Sri Lanka, Australia, Cambodia, Tanzania). The GF-B/V model discriminated bacterial from viral infection in the discovery cohort an area under the receiver operator curve (AUROC) of 0.93. Validation in an independent cohort demonstrated the GF-B/V model had an AUROC of 0.84 (95% CI 0.76-0.90) with overall accuracy of 81.6% (95% CI 72.7-88.5). Performance did not vary with age, demographics, or site. Host transcriptional response diagnostics distinguish bacterial and viral illness across global sites with diverse endemic pathogens.

Virus-specific and shared gene expression signatures in immune cells after vaccination in response to influenza and vaccinia stimulation

Background

In the vaccine era, individuals receive multiple vaccines in their lifetime. Host gene expression in response to antigenic stimulation is usually virus-specific; however, identifying shared pathways of host response across a wide spectrum of vaccine pathogens can shed light on the molecular mechanisms/components which can be targeted for the development of broad/universal therapeutics and vaccines.

Method

We isolated PBMCs, monocytes, B cells, and CD8+ T cells from the peripheral blood of healthy donors, who received both seasonal influenza vaccine (within <1 year) and smallpox vaccine (within 1 - 4 years). Each of the purified cell populations was stimulated with either influenza virus or vaccinia virus. Differentially expressed genes (DEGs) relative to unstimulated controls were identified for each in vitro viral infection, as well as for both viral infections (shared DEGs). Pathway enrichment analysis was performed to associate identified DEGs with KEGG/biological pathways.

Results

We identified 2,906, 3,888, 681, and 446 DEGs in PBMCs, monocytes, B cells, and CD8+ T cells, respectively, in response to influenza stimulation. Meanwhile, 97, 120, 20, and 10 DEGs were identified as gene signatures in PBMCs, monocytes, B cells, and CD8+ T cells, respectively, upon vaccinia stimulation. The majority of DEGs identified in PBMCs were also found in monocytes after either viral stimulation. Of the virus-specific DEGs, 55, 63, and 9 DEGs occurred in common in PBMCs, monocytes, and B cells, respectively, while no DEGs were shared in infected CD8+ T cells after influenza and vaccinia. Gene set enrichment analysis demonstrated that these shared DEGs were over-represented in innate signaling pathways, including cytokine-cytokine receptor interaction, viral protein interaction with cytokine and cytokine receptor, Toll-like receptor signaling, RIG-I-like receptor signaling pathways, cytosolic DNA-sensing pathways, and natural killer cell mediated cytotoxicity.

Conclusion

Our results provide insights into virus-host interactions in different immune cells, as well as host defense mechanisms against viral stimulation. Our data also highlights the role of monocytes as a major cell population driving gene expression in ex vivo PBMCs in response to viral stimulation. The immune response signaling pathways identified in this study may provide specific targets for the development of novel virus-specific therapeutics and improved vaccines for vaccinia and influenza. Although influenza and vaccinia viruses have been selected in this study as pathogen models, this approach could be applicable to other pathogens.

Benchmarking transcriptional host response signatures for infection diagnosis

Identification of host transcriptional response signatures has emerged as a new paradigm for infection diagnosis. For clinical applications, signatures must robustly detect the pathogen of interest without cross-reacting with unintended conditions. To evaluate the performance of infectious disease signatures, we developed a framework that includes a compendium of 17,105 transcriptional profiles capturing infectious and non-infectious conditions and a standardized methodology to assess robustness and cross-reactivity. Applied to 30 published signatures of infection, the analysis showed that signatures were generally robust in detecting viral and bacterial infections in independent data. Asymptomatic and chronic infections were also detectable, albeit with decreased performance. However, many signatures were cross-reactive with unintended infections and aging. In general, we found robustness and cross-reactivity to be conflicting objectives, and we identified signature properties associated with this trade-off. The data compendium and evaluation framework developed here provide a foundation for the development of signatures for clinical application. A record of this paper's transparent peer review process is included in the supplemental information.

Toward a more generalizable blood RNA signature for bacterial and viral infections

Host-response profiles can discriminate different infections. A new 8-gene blood RNA signature to discriminate bacterial and viral infections extends our focus hitherto on the case mix from the US and Europe to include that of low- and middle-income countries.1 Challenges remain.

A robust host-response-based signature distinguishes bacterial and viral infections across diverse global populations

Limited sensitivity and specificity of current diagnostics lead to the erroneous prescription of antibiotics. Host-response-based diagnostics could address these challenges. However, using 4,200 samples across 69 blood transcriptome datasets from 20 countries from patients with bacterial or viral infections representing a broad spectrum of biological, clinical, and technical heterogeneity, we show current host-response-based gene signatures have lower accuracy to distinguish intracellular bacterial infections from viral infections than extracellular bacterial infections. Using these 69 datasets, we identify an 8-gene signature to distinguish intracellular or extracellular bacterial infections from viral infections with an area under the receiver operating characteristic curve (AUROC) > 0.91 (85.9% specificity and 90.2% sensitivity). In prospective cohorts from Nepal and Laos, the 8-gene classifier distinguished bacterial infections from viral infections with an AUROC of 0.94 (87.9% specificity and 91% sensitivity). The 8-gene signature meets the target product profile proposed by the World Health Organization and others for distinguishing bacterial and viral infections.

Using Precision Medicine for the Diagnosis and Treatment of Viral Pneumonia

The COVID-19 pandemic has drawn considerable attention to viral pneumonia from clinicians, public health authorities, and the general public. With dozens of viruses able to cause pneumonia in humans, differentiating viral from bacterial pneumonia can be very challenging in clinical practice using traditional diagnostic methods. Precision medicine is a medical model in which decisions, practices, interventions, and therapies are adapted to the individual patient on the basis of their predicted response or risk of disease. Precision medicine approaches hold promise as a way to improve outcomes for patients with viral pneumonia. This review describes the latest advances in the use of precision medicine for diagnosing and treating viral pneumonia in adults and discusses areas where further research is warranted.

Antibiotic stewardship in the era of precision medicine

Antimicrobial resistance (AMR) continues to spread at an alarming rate worldwide. Novel approaches are needed to mitigate its deleterious impact on antibiotic efficacy. Antibiotic stewardship aims to promote the appropriate use of antibiotics through evidence-based interventions. One paradigm is precision medicine, a medical model in which decisions, practices, interventions, and therapies are adapted to the individual patient based on their predicted response or risk of disease. Precision medicine approaches hold promise as a way to improve outcomes for patients with myriad illnesses, including infections such as bacteraemia and pneumonia. This review describes the latest advances in precision medicine as they pertain to antibiotic stewardship, with an emphasis on hospital-based antibiotic stewardship programmes. The impact of the COVID-19 pandemic on AMR and antibiotic stewardship, gaps in the scientific evidence, and areas for further research are also discussed.