Comparing tuberculosis gene signatures in malnourished individuals using the TBSignatureProfiler

Comparison of gene set scoring methods for reproducible evaluation of tuberculosis gene signatures

BACKGROUND

Blood-based transcriptional gene signatures for tuberculosis (TB) have been developed with potential use to diagnose disease. However, an unresolved issue is whether gene set enrichment analysis of the signature transcripts alone is sufficient for prediction and differentiation or whether it is necessary to use the original model created when the signature was derived. Intra-method comparison is complicated by the unavailability of original training data and missing details about the original trained model. To facilitate the utilization of these signatures in TB research, comparisons between gene set scoring methods cross-data validation of original model implementations are needed.

METHODS

We compared the performance of 19 TB gene signatures across 24 transcriptomic datasets using both rrebuilt original models and gene set scoring methods. Existing gene set scoring methods, including ssGSEA, GSVA, PLAGE, Singscore, and Zscore, were used as alternative approaches to obtain the profile scores. The area under the ROC curve (AUC) value was computed to measure performance. Correlation analysis and Wilcoxon paired tests were used to compare the performance of enrichment methods with the original models.

RESULTS

For many signatures, the predictions from gene set scoring methods were highly correlated and statistically equivalent to the results given by the original models. In some cases, PLAGE outperformed the original models when considering signatures' weighted mean AUC values and the AUC results within individual studies.

CONCLUSION

Gene set enrichment scoring of existing gene sets can distinguish patients with active TB disease from other clinical conditions with equivalent or improved accuracy compared to the original methods and models. These data justify using gene set scoring methods of published TB gene signatures for predicting TB risk and treatment outcomes, especially when original models are difficult to apply or implement.

The sound of silent RNA in tuberculosis and the lncRNA role on infection

Tuberculosis (TB) is one of the leading causes of death worldwide, and Diabetes Mellitus is one of the major comorbidities (TB/DM) associated with the disease. A total of 103 differentially expressed ncRNAs have been identified in the TB and TB/DM comparisons. A machine learning algorithm was employed to identify the most informative lncRNAs: ADM-DT, LINC02009, LINC02471, SOX2-OT, and GK-AS1. These lncRNAs presented substantial accuracy in classifying TB from HC (AUCs >0.85) and TB/DM from HC (AUCs >0.90) in the other three countries. Genes with significant correlations with the five lncRNAs enriched common pathways in Brazil and India for both TB and TB/DM. This suggests that lncRNAs play an important role in the regulation of genes related to the TB immune response.

Comparison of gene set scoring methods for reproducible evaluation of multiple tuberculosis gene signatures

Rationale

Many blood-based transcriptional gene signatures for tuberculosis (TB) have been developed with potential use to diagnose disease, predict risk of progression from infection to disease, and monitor TB treatment outcomes. However, an unresolved issue is whether gene set enrichment analysis (GSEA) of the signature transcripts alone is sufficient for prediction and differentiation, or whether it is necessary to use the original statistical model created when the signature was derived. Intra-method comparison is complicated by the unavailability of original training data, missing details about the original trained model, and inadequate publicly-available software tools or source code implementing models. To facilitate these signatures' replicability and appropriate utilization in TB research, comprehensive comparisons between gene set scoring methods with cross-data validation of original model implementations are needed.

Objectives

We compared the performance of 19 TB gene signatures across 24 transcriptomic datasets using both re-rebuilt original models and gene set scoring methods to evaluate whether gene set scoring is a reasonable proxy to the performance of the original trained model. We have provided an open-access software implementation of the original models for all 19 signatures for future use.

Methods

We considered existing gene set scoring and machine learning methods, including ssGSEA, GSVA, PLAGE, Singscore, and Zscore, as alternative approaches to profile gene signature performance. The sample-size-weighted mean area under the curve (AUC) value was computed to measure each signature's performance across datasets. Correlation analysis and Wilcoxon paired tests were used to analyze the performance of enrichment methods with the original models.

Measurement and Main Results

For many signatures, the predictions from gene set scoring methods were highly correlated and statistically equivalent to the results given by the original diagnostic models. PLAGE outperformed all other gene scoring methods. In some cases, PLAGE outperformed the original models when considering signatures' weighted mean AUC values and the AUC results within individual studies.

Conclusion

Gene set enrichment scoring of existing blood-based biomarker gene sets can distinguish patients with active TB disease from latent TB infection and other clinical conditions with equivalent or improved accuracy compared to the original methods and models. These data justify using gene set scoring methods of published TB gene signatures for predicting TB risk and treatment outcomes, especially when original models are difficult to apply or implement.

A protein signature associated with active tuberculosis identified by plasma profiling and network-based analysis

Annually, approximately 10 million people are diagnosed with active tuberculosis (TB), and 1.4 million die of the disease. If left untreated, each person with active TB will infect 10-15 new individuals. The lack of non-sputum-based diagnostic tests leads to delayed diagnoses of active pulmonary TB cases, contributing to continued disease transmission. In this exploratory study, we aimed to identify biomarkers associated with active TB. We assessed the plasma levels of 92 proteins associated with inflammation in individuals with active TB (n = 20), latent TB (n = 14), or healthy controls (n = 10). Using co-expression network analysis, we identified one module of proteins with strong association with active TB. We removed proteins from the module that had low abundance or were associated with non-TB diseases in published transcriptomic datasets, resulting in a 12-protein plasma signature that was highly enriched in individuals with pulmonary and extrapulmonary TB and was further associated with disease severity.

Malnutrition leads to increased inflammation and expression of tuberculosis risk signatures in recently exposed household contacts of pulmonary tuberculosis

Background

Most individuals exposed to Mycobacterium tuberculosis (Mtb) develop latent tuberculosis infection (LTBI) and remain at risk for progressing to active tuberculosis disease (TB). Malnutrition is an important risk factor driving progression from LTBI to TB. However, the performance of blood-based TB risk signatures in malnourished individuals with LTBI remains unexplored. The aim of this study was to determine if malnourished and control individuals had differences in gene expression, immune pathways and TB risk signatures.

Methods

We utilized data from 50 tuberculin skin test positive household contacts of persons with TB - 18 malnourished participants (body mass index [BMI] < 18.5 kg/m2) and 32 controls (individuals with BMI ≥ 18.5 kg/m2). Whole blood RNA-sequencing was conducted to identify differentially expressed genes (DEGs). Ingenuity Pathway Analysis was applied to the DEGs to identify top canonical pathways and gene regulators. Gene enrichment methods were then employed to score the performance of published gene signatures associated with progression from LTBI to TB.

Results

Malnourished individuals had increased activation of inflammatory pathways, including pathways involved in neutrophil activation, T-cell activation and proinflammatory IL-1 and IL-6 cytokine signaling. Consistent with known association of inflammatory pathway activation with progression to TB disease, we found significantly increased expression of the RISK4 (area under the curve [AUC] = 0.734) and PREDICT29 (AUC = 0.736) progression signatures in malnourished individuals.

Conclusion

Malnourished individuals display a peripheral immune response profile reflective of increased inflammation and a concomitant increased expression of risk signatures predicting progression to TB. With validation in prospective clinical cohorts, TB risk biomarkers have the potential to identify malnourished LTBI for targeted therapy.

Comprehensive lipid and lipid-related gene investigations of host immune responses to characterize metabolism-centric biomarkers for pulmonary tuberculosis

Despite remarkable success in the prevention and treatment of tuberculosis (TB), it remains one of the most devastating infectious diseases worldwide. Management of TB requires an efficient and timely diagnostic strategy. In this study, we comprehensively characterized the plasma lipidome of TB patients, then selected candidate lipid and lipid-related gene biomarkers using a data-driven, knowledge-based framework. Among 93 lipids that were identified as potential biomarker candidates, ether-linked phosphatidylcholine (PC O–) and phosphatidylcholine (PC) were generally upregulated, while free fatty acids and triglycerides with longer fatty acyl chains were downregulated in the TB group. Lipid-related gene enrichment analysis revealed significantly altered metabolic pathways (e.g., ether lipid, linolenic acid, and cholesterol) and immune response signaling pathways. Based on these potential biomarkers, TB patients could be differentiated from controls in the internal validation (random forest model, area under the curve [AUC] 0.936, 95% confidence interval [CI] 0.865–0.992). PC(O-40:4), PC(O-42:5), PC(36:0), and PC(34:4) were robust biomarkers able to distinguish TB patients from individuals with latent infection and healthy controls, as shown in the external validation. Small changes in expression were identified for 162 significant lipid-related genes in the comparison of TB patients vs. controls; in the random forest model, their utilities were demonstrated by AUCs that ranged from 0.829 to 0.956 in three cohorts. In conclusion, this study introduced a potential framework that can be used to identify and validate metabolism-centric biomarkers.

Transcriptomics for child and adolescent tuberculosis

Tuberculosis (TB) in humans is caused by Mycobacterium tuberculosis (Mtb). It is estimated that 70 million children (<15 years) are currently infected with Mtb, with 1.2 million each year progressing to disease. Of these, a quarter die. The risk of progression from Mtb infection to disease and from disease to death is dependent on multiple pathogen and host factors. Age is a central component in all these transitions. The natural history of TB in children and adolescents is different to adults, leading to unique challenges in the development of diagnostics, therapeutics, and vaccines. The quantification of RNA transcripts in specific cells or in the peripheral blood, using high-throughput methods, such as microarray analysis or RNA-Sequencing, can shed light into the host immune response to Mtb during infection and disease, as well as understanding treatment response, disease severity, and vaccination, in a global hypothesis-free manner. Additionally, gene expression profiling can be used for biomarker discovery, to diagnose disease, predict future disease progression and to monitor response to treatment. Here, we review the role of transcriptomics in children and adolescents, focused mainly on work done in blood, to understand disease biology, and to discriminate disease states to assist clinical decision-making. In recent years, studies with a specific pediatric and adolescent focus have identified blood gene expression markers with diagnostic or prognostic potential that meet or exceed the current sensitivity and specificity targets for diagnostic tools. Diagnostic and prognostic gene expression signatures identified through high-throughput methods are currently being translated into diagnostic tests.

Development and Validation of a Parsimonious Tuberculosis Gene Signature Using the digital NanoString nCounter Platform

BACKGROUND

Blood-based biomarkers for diagnosing active tuberculosis (TB), monitoring treatment response, and predicting risk of progression to TB disease have been reported. However, validation of the biomarkers across multiple independent cohorts is scarce. A robust platform to validate TB biomarkers in different populations with clinical end points is essential to the development of a point-of-care clinical test. NanoString nCounter technology is an amplification-free digital detection platform that directly measures mRNA transcripts with high specificity. Here, we determined whether NanoString could serve as a platform for extensive validation of candidate TB biomarkers.

METHODS

The NanoString platform was used for performance evaluation of existing TB gene signatures in a cohort in which signatures were previously evaluated on an RNA-seq dataset. A NanoString codeset that probes 107 genes comprising 12 TB signatures and 6 housekeeping genes (NS-TB107) was developed and applied to total RNA derived from whole blood samples of TB patients and individuals with latent TB infection (LTBI) from South India. The TBSignatureProfiler tool was used to score samples for each signature. An ensemble of machine learning algorithms was used to derive a parsimonious biomarker.

RESULTS

Gene signatures present in NS-TB107 had statistically significant discriminative power for segregating TB from LTBI. Further analysis of the data yielded a NanoString 6-gene set (NANO6) that when tested on 10 published datasets was highly diagnostic for active TB.

CONCLUSIONS

The NanoString nCounter system provides a robust platform for validating existing TB biomarkers and deriving a parsimonious gene signature with enhanced diagnostic performance.

A 10-gene biosignature of tuberculosis treatment monitoring and treatment outcome prediction

The clinical utility of blood transcriptomic biosignatures for the treatment monitoring and outcome prediction of tuberculosis (TB) remains limited. In this study, we aimed to discover and validate biomarkers for pulmonary TB treatment monitoring and outcome prediction based on kinetic responses of gene expression during treatment. In particular, differentially expressed genes (DEGs) were identified by time-series comparison. Subsequently, DEGs with the monotonic expression alterations during the treatment were selected. Ten consistently down-regulated genes (CD274, KIF1B, IL15, TLR1, TLR5, FCGR1A, GBP1, NOD2, GBP2, EGF) exhibited significant potential in treatment monitoring, demonstrated via biological and technical validation. Additionally, the biosignature showed potential in predicting the cured versus relapsed patients. Furthermore, the biosignature could be utilized for TB diagnosis, latent tuberculosis infection/active TB differential diagnosis, and risk of progression to active TB. Benchmarking analysis of the 10-gene biosignature with other biosignatures showed equivalent performance in tested data sets. In conclusion, we established a 10-gene transcriptomic biosignature that represents the kinetic responses of TB treatment. Subsequent studies are warranted to validate, refine and translate the biosignature into a precise assay to assist clinical decisions in a broad spectrum of TB management.

Food for thought: addressing undernutrition to end tuberculosis

Tuberculosis is the leading cause of deaths from an infectious disease worldwide. WHO's End TB Strategy is falling short of several 2020 targets. Undernutrition is the leading population-level risk factor for tuberculosis. Studies have consistently found that undernutrition is associated with increased tuberculosis incidence, increased severity, worse treatment outcomes, and increased mortality. Modelling studies support implementing nutritional interventions for people living with tuberculosis and those at risk of tuberculosis disease to ensure the success of the End TB Strategy. In this Personal View, we highlight nutrition-related immunocompromisation, implications of undernutrition for tuberculosis treatment and prevention, the role of nutritional supplementation, pharmacokinetics and pharmacodynamics of antimycobacterial medications in undernourished people with tuberculosis, and the role of social protection interventions in addressing undernutrition as a tuberculosis risk factor. To catalyse action on this insufficiently addressed accelerant of the global tuberculosis epidemic, research should be prioritised to understand the immunological pathways that are impaired by nutrient deficiencies, develop tools to diagnose clinical and subclinical tuberculosis in people who are undernourished, and understand how nutritional status affects the efficacy of tuberculosis vaccine and therapy. Through primary research, modelling, and implementation research, policy change should also be accelerated, particularly in countries with a high burden of tuberculosis.