Multiplexed cytokine profiling identifies diagnostic signatures for latent tuberculosis and reactivation risk stratification

Abstract

Active tuberculosis (TB) is caused by Mycobacterium tuberculosis (Mtb) bacteria and is characterized by multiple phases of infection, leading to difficulty in diagnosing and treating infected individuals. Patients with latent tuberculosis infection (LTBI) can reactivate to the active phase of infection following perturbation of the dynamic bacterial and immunological equilibrium, which can potentially lead to further Mtb transmission. However, current diagnostics often lack specificity for LTBI and do not inform on TB reactivation risk. We hypothesized that immune profiling readily available QuantiFERON-TB Gold Plus (QFT) plasma supernatant samples could improve LTBI diagnostics and infer risk of TB reactivation. We applied a whispering gallery mode, silicon photonic microring resonator biosensor platform to simultaneously quantify thirteen host proteins in QFT-stimulated plasma samples. Using machine learning algorithms, the biomarker concentrations were used to classify patients into relevant clinical bins for LTBI diagnosis or TB reactivation risk based on clinical evaluation at the time of sample collection. We report accuracies of over 90% for stratifying LTBI + from LTBI- patients and accuracies reaching over 80% for classifying LTBI + patients as being at high or low risk of reactivation. Our results suggest a strong reliance on a subset of biomarkers from the multiplexed assay, specifically IP-10 for LTBI classification and IL-10 and IL-2 for TB reactivation risk assessment. Taken together, this work introduces a 45-minute, multiplexed biomarker assay into the current TB diagnostic workflow and provides a single method capable of classifying patients by LTBI status and TB reactivation risk, which has the potential to improve diagnostic evaluations, personalize treatment and management plans, and optimize targeted preventive strategies in Mtb infections.

Fig 1. LTBI classification using absolute and normalized stimulated cytokine concentrations.

Random forest classification is represented as receiver operator characteristic (ROC) curves when using absolute (A) and normalized (B) stimulated biomarker concentrations as the input variables. The most important variables identified through the variable importance metrics using the full set of absolute (C) and normalized (D) biomarker concentrations were used to generate the reduced data models. The variable importance metrics for all input variables in both models are presented in the Supplemental Information.

Fig 2. Significantly different distributions of absolute cytokine values between LTBI positive and negative patients.

Target concentrations from stimulation conditions of TB1 (A), NIL (B), and MIT (C) that were significantly different between clinical populations of LTBI+ and LTBI− patients are shown. Median value of clinical population represented as a black line. Data visualized on a log scale. Wilcox-Mann-Whitney tests were completed to test significant differences in target concentrations between clinical population distributions. *  p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001, **** p ≤ 0.0001.

Fig 3. Significantly different normalized distributions between LTBI positive and negative patients.

Normalized conditions of TB1-NIL (A), TB1-MIT (B), and MIT-NIL (C) were significantly different between clinical populations of LTBI+ and LTBI− patients in some measured targets. Median value of clinical population represented as a black line. Wilcox-Mann-Whitney tests were completed to test for significant differences in normalized target concentrations between clinical population distributions. *  p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001, **** p ≤ 0.0001.

Fig 4. High-risk classification of LTBI positive patients using absolute and normalized stimulated cytokine concentrations.

Random forest classification is represented as receiver operator characteristic (ROC) curves when using absolute (A) and normalized (B) stimulated biomarker concentrations. The most important variables from the full data set were identified through the variable importance metrics of absolute (C) and normalized (D) biomarker concentrations. Reduced data models were generated using the top biomarkers presented here. The variable importance metrics for all input variables in both models are presented in the Supplemental Information.

Fig 5. Low-risk classification of LTBI positive patients using absolute and normalized stimulated cytokine concentrations.

Random forest classification is represented as receiver operator characteristic (ROC) curves when using absolute (A) and normalized (B) stimulated biomarker concentrations. The most important variables identified through the variable importance metrics using absolute (C) and normalized (D) biomarker concentrations were used to generate the reduced data models. The variable importance metrics for all input variables in both models are presented in the supplemental information.

Fig 6. Significantly different absolute concentration distributions between high-risk and low-risk LTBI positive patients.

Targets under stimulation conditions of TB1 (A), NIL (B), and MIT (C) that were significantly different between clinical populations of high and low risk LTBI+ patients are shown. Median value of clinical population represented as a black line. Data visualized on log scale. Wilcox-Mann-Whitney tests were completed to test for significant differences in normalized target concentrations between clinical population distributions. *  p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001, **** p ≤ 0.0001.