Diverse impacts of different rpoB mutations on the anti-tuberculosis efficacy of capreomycin.

BACKGROUND

Since the discovery of streptomycin in the 1940s, more than a dozen drugs have been continuously introduced into tuberculosis (TB) therapy. However, limited attention has been paid to the collateral effects of drug resistance evolution in Mycobacterium tuberculosis (Mtb). Recently, we observed a clear discordance between the capreomycin (CAP) susceptibility of rifampicin-resistant (RR) Mtb clinical isolates and the adverse outcomes associated with CAP treatment, indicating potential collateral effects between rpoB mutations and CAP. To explore this relationship, we integrated clinical isolate data, experimental evolution data, phenotypic data, sequencing data, and genome-wide association studies (GWAS).

METHODS

We analysed the correlations between CAP resistance and rpoB mutations at various loci based on phenotypic drug susceptibility testing (pDST) profiles and rpoB sequencing data from 565 RR Mtb isolates collected in southwestern China. To validate the clinical observations, we screened RR mutants of Mtb H37Rv and conducted rpoB sequencing to characterise the mutation sites. Additionally, we constructed various rpoB mutants in Mycobacterium smegmatis (Ms). We then examined the impact of these mutations on the efficacy of CAP through minimum inhibitory concentration (MIC) tests and time-kill assays in both Mtb and Ms rpoB mutants. Furthermore, we investigated the influence of three major rpoB mutations on the frequency of occurrence of rrs A1401G-associated with CAP resistance-using a GWAS of 607 Mtb genomes from a global dataset.

FINDINGS

By analysing 565 clinical isolates from southwestern China, we found that the CAP resistance in isolates with a single mutation at rpoB site 445 was significantly lower than in those with a single mutation at other sites (P < 0.05, Pearson chi-square test and Fisher exact test; odds ratio = 0.272). In contrast, the opposite trend was observed in isolates with a single mutation at rpoB site 435 (P < 0.001, Pearson chi-square test and Fisher exact test; odds ratio = 3.067). Subsequently, using laboratory-evolved RR mutants, we demonstrated that mutations at rpoB site 445 or site 441 enhanced the bactericidal effect of CAP. However, the opposite result was observed in mutants with mutations at rpoB site 435. Furthermore, we found that the occurrence frequency of the rrs A1401G mutation was significantly lower in clinical isolates with rpoB mutations at site 445, but significantly higher in those with mutations at site 435.

INTERPRETATION

Although rpoB mutations in Mtb did not affect the MIC of CAP, they influenced its bactericidal effect, highlighting the need for time-kill assays when investigating collateral effects. Different rpoB mutations may exert diverse impacts on the bactericidal effect of CAP-or CAP tolerance-underscoring the complexity of collateral effects and supporting the use of targeted sequencing in the molecular diagnosis of RR Mtb. As RNA polymerase plays a central role in bacterial RNA transcription, it regulates most metabolic processes in Mtb. Thus, different rpoB mutations may elicit distinct gene expression profiles upon CAP treatment, a hypothesis warranting further investigation. Additional clinical studies are needed to verify whether the adverse outcomes of CAP treatment are associated with infections caused by strains harbouring rpoB mutations at site 435. If so, such outcomes could be mitigated through rational drug regimens guided by precise molecular diagnosis. This study provides insights into the collateral effects of drug resistance mutations and advances the case for precision medicine in treating infections caused by drug-resistant bacteria.

FUNDING

Funding for this study was provided by the National Key Research and Development Program of China (Grant no. 2021YFA1300901).