Benchtop nuclear magnetic resonance‐based metabolomic approach for the diagnosis of bovine tuberculosis. Issue 4 (19th November 2021)

• Record Type: Journal Article
• Authors: Ruiz‐Cabello, Jesús; Sevilla, Iker A.; Olaizola, Ekine; Bezos, Javier; Miguel‐Coello, Ana B.; Muñoz‐Mendoza, Marta; Beraza, Marta; Garrido, Joseba M.; Izquierdo‐García, Jose L.
• Is Part Of: Transboundary and emerging diseases. Volume 69:Issue 4(2022)
• Page Start: e859
• Page End: e870
• Publication Date: 2021-11-19
• Subjects: benchtop -- bovine -- diagnosis -- metabolomics -- NMR -- tuberculosis; Veterinary medicine -- Periodicals; 636.089
• DOI: 10.1111/tbed.14365
• Languages: English
• ISSNs: 1865-1674

Abstract

Abstract: Even though enormous efforts and control strategies have been implemented, bovine tuberculosis (TB) remains a significant source of health and socioeconomic concern. The standard method used in TB eradication programs for in vivo detection is the tuberculin skin test. However, the specificity of the tuberculin skin test is affected by infection with non‐tuberculous mycobacteria or by vaccination. Thus, some animals are not correctly diagnosed. This study aimed first to identify a plasma metabolic TB profile by high‐field (HF) nuclear magnetic resonance (NMR) spectroscopy and second measure this characteristic TB metabolic profile using low‐field benchtop (LF) NMR as an affordable molecular technology for TB diagnosis. Plasma samples from cattle diagnosed with TB (derivation set, n = 11), diagnosed with paratuberculosis (PTB, n = 10), PTB‐vaccinated healthy control (n = 10) and healthy PTB‐unvaccinated control (n = 10) were analyzed by NMR. Unsupervised Principal Component Analysis (PCA) was used to identify metabolic differences between groups. We identified 14 metabolites significantly different between TB and control animals. The second group of TB animals was used to validate the results (validation set, n = 14). Predictive models based on metabolic fingerprint acquired by both HF and LF NMR spectroscopy successfully identified TB versus control subjects (Area under the curve of Receiver Operating Characteristic over 0.92, in both models; Confidence Interval 0.77–1). In summary, plasma fingerprinting using HF and LF‐NMR differentiated TB subjects from uninfected animals, and PTB and PTB‐vaccinated subjects who may provide a TB‐false positive, highlighting the use of LF‐NMR‐based metabolomics as a complementary or alternative diagnostic tool to the current diagnostic methods.