Cross-Resistance between Clofazimine and Bedaquiline through Upregulation of MmpL5 in Mycobacterium tuberculosis

TEXT

Clofazimine, a riminophenazine, is a standard component of the multidrug therapy used for the treatment of leprosy and has contributed to the cure of 16 million patients. In recent years, due to the spread of multidrug-resistant Mycobacterium tuberculosis strains there has been renewed interest in the use of clofazimine for treatment of multidrug-resistant tuberculosis (1). Clofazimine has been elegantly demonstrated in M. tuberculosis to be a prodrug, which is reduced by NADH dehydrogenase (Ndh2), and then upon spontaneous reoxidation by O₂, to release reactive oxygen species (ROS) (2). As there is no specific target for ROS, resistance to clofazimine is rare and to our knowledge has never been characterized.

With the aim of further understanding the mechanism of action of resistance of clofazimine, we sought to determine the means by which M. tuberculosis strain H37Rv can develop resistance to the compound. Resistant mutants were selected by plating H37Rv (100 μl at an optical density at 600 nm [OD₆₀₀] of 1) on Middlebrook 7H11 medium containing clofazimine at concentrations only just greater (4×) than its MIC (solid clofazimine MIC, 0.25 μg/ml; isolation at 1 μg/ml). Following the passage of an isolated colony (H37Rv_CFZ-R1) in the absence of antibiotic pressure, resistance to clofazimine was maintained (4- to 8-fold), as determined using the resazurin reduction microplate assay (REMA) (3) (Fig. 1), suggesting the involvement of a genetic resistance mechanism. To evaluate the specificity of the mechanism of resistance, the susceptibility of H37Rv_CFZ-R1 was determined for a panel of antituberculosis drugs. While the MIC of H37Rv_CFZ-R1 for rifampin, PA-824, isoniazid, and moxifloxacin remained unchanged, marked cross-resistance was observed with bedaquiline (4- to 8-fold) and, to a much lesser extent, to the azole drug econazole (Fig. 1 and Table 1).

To identify the genetic mutation causing the clofazimine-bedaquiline cross-resistance, the genomes of H37Rv and H37Rv_CFZ-R1 were subjected to whole-genome sequencing using Illumina technology. Bioinformatics analysis (as described in reference 4) identified a nonsynonymous single nucleotide polymorphism in gene rv0678 (c189a leading to S63R) in H37Rv_CFZ-R1, while atpE contained no mutations. The presence of this mutation was confirmed independently by Sanger sequencing. A second de novo round of mutant isolation yielded five further clofazimine-resistant mutants. Sanger sequencing of rv0678 from these five isolates showed that four of them carried the same c189a mutation identified in H37Rv_CFZ-R1, and one (H37Rv_CFZ-R2) harbored a novel c400t mutation in rv0678 that results in a premature stop codon (R134stop).

Rv0678 is a transcriptional repressor with a winged helix DNA binding domain that has been shown to bind to a palindromic sequence located in the intergenic region between rv0678 and the neighboring genes, mmpS5 and mmpL5 (5). Numerous mutations in Rv0678 were found to lead to derepression of this locus with subsequent transcriptional upregulation of all three genes (rv0678, mmpS5, and mmpL5) (5). Here, using quantitative PCR (qPCR), we confirmed that H37Rv_CFZ-R1 carrying Rv0678_(S63R) displays 11.6- and 11.2-fold-increased expression of mmpL5 and rv0678, respectively, relative to wild-type H37Rv.

To confirm genetically the role of Rv0678_(S63R) in resistance to clofazimine and bedaquiline, rv0678_(S63R) was expressed from the hsp60 promoter in the integrative vector pMV261 in M. tuberculosis strain H37Rv. Drug susceptibility evaluation by REMA clearly showed that expression of Rv0678_(S63R) in H37Rv leads to partial resistance to both clofazimine and bedaquiline (Table 1). Partial, and not complete, resistance is likely because the chromosomal wild-type rv0678 gene is still expressed. Ectopic overexpression of Rv0678_(S63R) in H37Rv_CFZ-R1 had no further impact on drug susceptibility.

MmpL proteins (6) are multisubstrate efflux pumps that are part of the family of “resistance, nodulation, and cell division” (RND) proteins (7). MmpL5 is one of 13 such MmpL proteins found in M. tuberculosis, and together with MmpL4 has been shown to be involved in the export of siderophores (mycobactins and carboxymycobactins) for the bacterial acquisition of iron (8, 9). MmpL5 by itself has been found to be dispensable in M. tuberculosis (7), although removal of both MmpL5 and MmpL4 has a detrimental impact on both bacterial virulence and viability in vivo (8). Mutations in rv0678, with subsequent upregulation of mmpS5 and mmpL5, have been shown to lead to resistance to azoles (5), as is also seen in the present study with resistance to econazole (Fig. 1). Resistance to azoles was demonstrated to be through the active efflux of these compounds by MmpL5, suggesting that clofazimine and bedaquiline may also be substrates for this multisubstrate efflux pump. This hypothesis is also supported by data demonstrating that the general inhibitor of efflux, verapamil, potentiates bedaquiline activity in M. tuberculosis (10). Nevertheless, it is intriguing that MmpL5 seems to affect in particular compounds acting on or affected by the electron transport chain (azoles, clofazimine, and bedaquiline). In addition, bedaquiline has been shown to act synergistically with both pyrazinamide in vivo (11) and benzothiazinones in vitro (12, 13), and a possible source of this synergy may be through modulation of efflux pumps, such as MmpL5.

Bedaquiline is currently one of the most promising novel antituberculosis drug candidates (14). In previous work with bedaquiline-resistant H37Rv mutants isolated in vitro, it was found that 15 out of 53 resistant mutants carried mutations in atpE, the gene that encodes the target subunit of bedaquiline in ATP synthase (15, 16). Nonetheless, the remaining 38 resistant mutants did not harbor mutations in atpE, and their resistance remains unexplained. These “off-target” genetic mutants typically showed a 2- to 16-fold level of increased resistance, a similar magnitude to that seen here in strains with mutations in rv0678. In the absence of data explaining non-atpE-mediated bedaquiline resistance, it remains to be determined whether mutations in rv0678 are involved and whether these mutants display cross-resistance to clofazimine. The Global Alliance for TB Drug Development is currently conducting clinical trials to evaluate the activity of both clofazimine and bedaquiline in combination regiments for the treatment of tuberculosis. To date, very little is known about the potential sources of drug resistance in such regimens; however, mutations in rv0678 should be taken into consideration as a possible confounding factor. An additional intriguing point is the absence of the rv0678-mmpS5 mmpL5 locus in Mycobacterium leprae, and this may explain why leprosy is so effectively treated with clofazimine and resistance has never been encountered. In light of these findings, Rv0678 will be given the name MmpR5 (Mycobacterial membrane protein repressor).