In vitro selected mutants of M. tuberculosis do not reflect the in vivo mechanism of isoniazid resistance

Abstract number: P1950

Bergval I., Schuitema A., Klatser P., Anthony R.

Background: Multidrug-resistant tuberculosis (MDR-TB) is by definition resistant to at least isoniazid (INH) and rifampicin (RIF), the two most commonly used drugs against M. tuberculosis (MTB). Monoresistance to INH is much more common in clinical isolates than monoresistance to RIF. In addition, the in vitro rate of INH resistance is typically hundreds of times higher than the rate for RIF resistance. It is generally assumed that the high in vitro mutation frequencies explain the high prevalence of INH resistance in vivo.

We studied the in vitro mechanisms of INH resistance, in part to determine the preferred genetic routes to MDR under various conditions. Here we discuss our findings on INH resistance mechanisms in vitro.

Methods: We determined the frequencies and rates of spontaneous mutations conferring INH resistance under various conditions, using well-characterised MTB laboratory strains. Mutation frequencies were determined by counting the proportion of resistant mutants on solid medium and mutation rates were determined by Luria and Delbrück fluctuation assays (po-method). For each experiment and condition INH-resistant mutants were picked randomly and characterised by PCR, sequencing or MLPA.

Results: We found that the INH resistance rates and frequencies were approximately two logs higher (10^-6) than those previously measured for RIF resistance (10^-8). However, the katG gene was partially or completely deleted in the majority (30–90%, depending on the strain and condition) of the INH-resistant mutants. In contrast, INH-resistance in vivo is almost always due to point mutations in katG-315 or inhA(-15).

Attempts to select against katG deletion mutants, in favour of clinically relevant INH-resistant mutants, by selecting under oxidative conditions, were unsuccessful.

Conclusion: The high prevalence of INH-resistance in vivo has been attributed to the high rate of mutations conferring resistance to this drug in vitro. Based on our results and the scarce literature available, we feel that this is at best a simplification as the majority of the in vitro mutations are rarely seen in clinical isolates. Moreover, KatG is the only active defence against oxidative assault in MTB, but the prevalence of katG deletion mutants was not significantly reduced under oxidative conditions.

These results indicate that additional research into the emergence of INH resistance is needed and that assumptions based on the observed INH resistance rates in vitro may be unwarranted.