Mycobacterium tuberculosis cell wall contains antigenic glycolipids: phenol-glycolipid (PGL), diacyltrehalose (DAT), triacyltrehalose (TAT), cord-factor (CF), and sulpholipid-I (SL-I). In the last decade, the usefulness of these antigens for the serodiagnosis of tuberculosis has been evaluated mainly using enzyme-linked immunosorbent assays (ELISA). Currently, there are no conclusive results about the utility of these glycolipidic antigens, because the results obtained by different groups are discrepant. In order to explain these discrepancies, we have investigated the methodological variations in the ELISAs used previously. Specifically, we have studied the following: the coating solvent, the optimum amount of glycolipid coated per well, the blocking agent, and the use of detergent (Tween 20) in the washing buffer. The most significant finding was that Tween 20 detaches PGL, DAT, TAT and SL-I from microtitre wells. However, Tween 20 does not remove CF from the wells. In addition, we have found that the best solvent for coating is n-hexane, that the optimum antigen coating concentration is 1000 ng/well, and that BSA and gelatin are equally effective blocking agents. We can therefore conclude that the use of Tween 20 as a detergent, and the lower antigen coating concentrations (100-200 ng/well), may well explain some of the discrepancies in previous studies. © 2001 Elsevier Science B.V.
- Tween 20