Mycobacterium tuberculosis embB codon 306 mutations confer moderately increased resistance to ethambutol in vitro and in vivo

Abstract

Ethambutol (EMB) is a major component of the first-line therapy of tuberculosis. Mutations in codon 306 of embB (embB306) were suggested as a major resistance mechanism in clinical isolates. To directly analyze the impact of individual embB306 mutations on EMB resistance, we used allelic exchange experiments to generate embB306 mutants of M. tuberculosis H37Rv. The level of EMB resistance conferred by particular mutations was measured in vitro and in vivo after EMB therapy by daily gavage in a mouse model of aerogenic tuberculosis. The wild-type embB306 ATG codon was replaced by embB306 ATC, ATA, or GTG, respectively. All of the obtained embB306 mutants exhibited a 2- to 4-fold increase in EMB MIC compared to the wild-type H37Rv. In vivo, the one selected embB306 GTG mutant required a higher dose of ethambutol to restrict its growth in the lung compared to wild-type H37Rv. These experiments demonstrate that embB306 point mutations enhance the EMB MIC in vitro to a moderate, but significant extent, and reduce the efficacy of EMB treatment in the animal model. We propose that conventional EMB susceptibility testing, in combination with embB306 genotyping, may guide dose adjustment to avoid clinical treatment failure in these low-level resistant strains.

Fig. 1.

Bacterial growth of wild-type H37Rv and embB306 GTG mutant 14-9 in EMB-treated mice compared to untreated control groups. C57BL/6 female mice were inoculated via aerosol with 100 CFU of M. tuberculosis H37Rv (solid bars) or embB306 GTG mutant 14-9 (striped bars). The bacterial burden of both strains (d1, n = 3) was determined in the lung 24 h after infection resulting in 2.43 CFU (mean, log₁₀) for H37Rv and 2.40 CFU (mean, log₁₀) for mutant 14-9. The EMB treatment started 7 days after infection. Different EMB dosages (6.25, 12.5, 25, or 50 mg/kg/day, respectively; increasing gray shades in columns indicate increases in dose per treatment group) were administered by gavage. Control mice received nonsupplemented sterile water. The absolute bacterial load in the lung (A), in the liver (B) and in the spleen (C) was determined at day 28 and 49 after infection (n = 5, mean CFU + the SD). *, P < 0.05 for comparisons between individual treatments versus nontreated, infected controls.