A multiplexed and miniaturized serological tuberculosis assay identifies antigens that discriminate maximally between TB and non-TB sera

• Published in: Journal of immunological methods 2005-06, Vol.301 (1), p.154-163
• Main Authors: Tong, Miao, Jacobi, Catharina E., van de Rijke, Frans M., Kuijper, Sjoukje, van de Werken, Sjaak, Lowary, Todd L., Hokke, Cornelis H., Appelmelk, Ben J., Nagelkerke, Nico J.D., Tanke, Hans J., van Gijlswijk, Rob P.M., Veuskens, Jacques, Kolk, Arend H.J., Raap, Anton K.
• Format: Article
• Language: English
• Subjects: Antibodies, Bacterial - immunology; Antibody Specificity; Antigens, Bacterial - immunology; Biological and medical sciences; Carbohydrate Conformation; Carbohydrate Sequence; Fundamental and applied biological sciences. Psychology; Fundamental immunology; Humans; Immune Sera - immunology; Microarray; Molecular immunology; Molecular Sequence Data; Multi-variant analysis; Multiplexing; Mycobacterium tuberculosis - immunology; Oligosaccharides - chemistry; Oligosaccharides - metabolism; Protein Array Analysis; ROC analysis; ROC Curve; Sensitivity and Specificity; Serologic Tests - methods; Serology; Techniques; Tuberculosis; Tuberculosis - blood; Tuberculosis - diagnosis; Tuberculosis - immunology; Tuberculosis - microbiology
• ISSN: 0022-1759; 1872-7905
• DOI: 10.1016/j.jim.2005.04.004

We have developed a multiplexed and miniaturized TB serological assay with the aim of identifying (combinations of) antigens that maximally discriminate between TB and non-TB patients. It features a microarray accommodating 54 TB antigens, less than 1 μl serum consumption and an indirect immunofluorescence detection protocol. With a panel of 20 TB and 80 non-TB sera we ranked combinations of TB antigens with respect to sensitivity and specificity of TB detection by means of logistic step-forward regression analysis. The highest-ranking TB antigen combination had an area-under-the-curve of the receiver–operator-characteristics (ROC) of 0.95. We also identified an antigen that on its own provided good specificity and sensitivity of TB detection (Ara 6–BSA; area-under-the-ROC curve: 0.90). These area-under-the-ROC curve values are exceptionally high for a serological TB assay. We conclude that TB antigen microarrays permit rapid identification of TB antigens that, either alone or in combination, discriminate maximally between TB and non-TB patients and that such identification provides an excellent starting point for developing point-of-care diagnostic assays.