Implications of drug-induced phenotypical resistance: Is isoniazid radicalizing M. tuberculosis?

Abstract

Background Tuberculosis treatment duration is long and does not guarantee eradication of infection. Shorter treatment regimens are a critical research objective to improve uptake and reduce the risk of relapse and bacterial resistance. The explanation for the need to continue treatment after patients are culture negative remains elusive. We have previously shown that the presence of lipid inclusions in mycobacterial cells is associated with an increase in antibiotic resistance. Aim We investigate the bactericidal effect of isoniazid and rifampicin on the expression of lipid inclusions and characterize the degree of the associated phenotypic antibiotic resistance to a range of anti-tuberculosis agents in current use. Methods Antibiotic killing effect for both M. tuberculosis and M. komossense were investigated by both hollow fiber bioreactor (HFS) studies and static time kill curve (STKC) experiments. Following STKC cultures were stained with resazurin, Sytox green and Nile red to establish their live/dead (resazurin positive/Sytox positive) and lipid inclusion status, respectively. In addition, M. komossense was studied in the hollow fiber bioreactor model (HFS) and exposed to isoniazid (H) and rifampicin (R). The MIC of current antituberculosis agents for cells from the treated hollow fiber experiments were tested. Results Antibiotic killing was similar for both species. For M. komossense; isoniazid was ineffective at the established MIC (1 mg/L) in the hollow fiber bioreactor but rifampicin reduced the viable count rapidly at MIC (0.4 mg/L). When the two drugs were combined at their respective MICs the killing effect was significant and greater than separately. Cells exposed to isoniazid (1x and 9x MIC) for 168 h showed considerable numbers of recoverable viable cells when compared with a combination of 1x MIC R & H where there were no viable cells detectable. For both drugs the number of lipid body positive cells increased over time and this effect was most pronounced for isoniazid and was associated with phenotypic resistance to multiple anti-tuberculosis drugs. Conclusion Our results showed that isoniazid is a potent stimulator of lipid body accumulation, culture persistence, and phenotypic resistance to multiple anti-tuberculosis drugs. These findings emphasize the importance of understanding mechanisms of drug-drug interactions and phenotypic resistance in regimen building.