Identification of potential biomarkers of disease progression in bovine tuberculosis

Abstract

Bovine tuberculosis (bTB) remains an important animal and zoonotic disease in many countries. The diagnosis of bTB is based on tuberculin skin test and IFN-γ release assays (IGRA). Positive animals are separated from the herd and sacrificed. The cost of this procedure is difficult to afford for developing countries with high prevalence of bTB; therefore, the improvement of diagnostic methods and the identification of animals in different stages of the disease will be helpful to control the infection. To identify biomarkers that can discriminate between tuberculin positive cattle with and without tuberculosis lesions (ML+ and ML−, respectively), we assessed a group of immunological parameters with three different classification methods: lineal discriminant analysis (LDA), quadratic discriminant analysis (QDA) and K nearest neighbors (k-nn). For this purpose, we used data from 30 experimentally infected cattle. All the classifiers (LDA, QDA and k-nn) selected IL-2 and IL-17 as the most discriminatory variables. The best classification method was LDA using IL-17 and IL-2 as predictors. The addition of IL-10 to LDA improves the performance of the classifier to discriminate ML-individuals (93.3% vs. 86.7%). Thus, the expression of IL-17, IL-2 and, in some cases, IL-10 would serve as an additional tool to study disease progression in herds with a history of bTB.