Several recent studies have demonstrated that many efflux pumps are upregulated in RIF-resistant clinical Mtb isolates [17,18,19]. Which efflux pumps play a central role in the regulation of RIF resistance in Mtb, however, remains uncertain, as does the relationship between particular rpoB gene mutation sites and the contribution of efflux pump activity to overall levels of RIF resistance. There is thus an urgent need for further studies aimed at probing the mechanistic roles played by efflux pumps in clinical RIF-resistant Mtb isolates so as to provide novel therapeutic options to patients with tuberculosis.

All of these clinical isolates harbored known rpoB resistance mutations. In an effort to clarify the extent to which particular efflux pumps contribute to the drug resistance of these strains, the impact of the efflux pump inhibitor VP on these strains was assessed. In response to VP, the MIC values for these strains declined by 2- to 32-fold. These findings are consistent with several other reports documenting a synergistic interaction between VP and RIF [15, 20,21,22,23]. RIF-resistant clinical isolates carrying the same rpoB gene mutation (e.g. Ser450Leu) exhibited varying degrees of RIF resistance (MICs ranging from 8 to 256 µg/mL), with this variability potentially being related to the numbers of overexpressed efflux pump genes, the magnitude of this overexpression, and the activity of the encoded efflux pumps. This is further supported by the finding that the MIC value for clinical isolates exposed to the efflux pump inhibitor VP did not exceed 16 µg/mL and that such treatment reduced differences in MIC values among strains. Prior reports have demonstrated that the emergence of RIF resistance is a continuous process. As efflux pump activity can maintain low levels of drug within Mtb cells, this can enable these pathogens to survive more readily in response to sublethal drug doses, thereby facilitating the emergence of drug resistance [24,25,26,27,28]. Based on the decreases in MIC values for these rpoB mutant strains following VP exposure, the contributions of efflux pumps to drug resistance appear to not only persist following the mutation of drug target genes in Mtb cells, but to actually play an even greater role in drug resistance.

When qPCR analyses of cultures exposed to RIF drug pressure were conducted, 34 of the RIF-resistant clinical isolates (all except for M21) harboring rpoB gene mutations exhibited the upregulation of anywhere from 1 to 13 of the following efflux pump genes: Rv2333c, Rv2937, Rv2938, Rv0451c, Rv0450c, Rv0677c, Rv2846c, Rv1819c, Rv0933, Rv0191, Rv1258c, Rv1410c and Rv2459. Each of these genes was overexpressed by more than twofold in a minimum of two of the analyzed isolates (Table 1). Li et al. previously analyzed MDR Mtb isolates and found that under similar RIF pressure (1/2 MIC), the Rv2937, Rv2846c, Rv1258c, Rv1410c, Rv2333c, Rv2459, and Rv0933 genes were upregulated greater than two-fold [16], in line with the present results. Although their study did not include any additional efflux pump genes, a separate study revealed that under pressure imposed by a 1/4 MIC dose of RIF, Rv0191, Rv2938, and Rv1819c were upregulated by more than twofold [15].

Li et al. also evaluated RIF monotolerant isolates, and observed no instances of Rv0191 gene overexpression by more than fourfold, whereas in the present study, 9 of the analyzed RIF-resistant isolated presented with greater than fourfold Rv0191 gene overexpression. This difference may be related to the use of sublethal RIF concentrations, with higher RIF doses potentially inducing higher levels of efflux pump expression. Isolates with different resistance phenotypes may also present with differences in efflux pump expression, underscoring a need for further research. Following RIF overexpression, efflux pump gene overexpression may not occur in a random manner. For example, Rv0677c and Rv0191 were the most frequently overexpressed efflux pump genes in the present study. When van der Heijden et al. conducted an expression quantitative trait locus (eQTL) analysis of fluoroquinolone-resistant clinical Mtb isolates, they identified Rv0191 as a potential mediator of fluoroquinolone efflux [29]. Zhang et al. additionally employed Mtb proteome microarrays to analyze the binding activity of the active component of pyrazinamide following its pyrazinamidase-mediated conversion, revealing that Rv0191 may facilitate the active efflux of this compound [30]. Xu et al. additionally determined that overexpressing Rv0677c and Rv0676c in the H37Rv Mtb strain can confer resistance to Bedaquinoline and Clofazimine [31].

However, little research to date has focused on the role that Rv0677c and Rv0191 play as mediators of RIF resistance in clinical Mtb isolates. Based on these past results and the present findings, Rv0667c and Rv0191 may function as broad-spectrum mediators of the active efflux of multiple drugs from Mtb cells. There have also been several prior studies suggesting that Rv0678 gene mutations mediate MmpL5-mmps5 upregulation [32,33,34,35]. In this study, however, 22 isolates exhibited significant (> fourfold) mmps5 overexpression following RIF exposure, but whole genome sequencing analyses of these isolates revealed no mutations in the Rv0678 gene. These Rv0678 gene mutations thus appear not to be necessary for MmpL5-mmps5 upregulation. Here, no analyzed isolates presented with greater than fourfold upregulation of Rv1410c, with just two exhibiting > twofold upregulation of this gene. Kardan-Yamchi et al. determined that in the absence of antibiotic pressure, certain efflux pump genes may be upregulated [36]. They specifically evaluated expression levels for four efflux pump genes in 31 clinical Mtb isolates and detected Rv1410c overexpression in just 7 out of 21 analyzed M/XDR-TB isolates.

Consistent with our data, Long et al. also observed the upregulation of certain efflux pump genes in some drug-resistant Mtb strains even without drug treatment [37]. The RIF-resistant clinical isolate M21 harboring the His445Tyr rpoB gene mutation did not exhibit the > twofold upregulation of any efflux pump genes following RIF exposure, potentially because these genes were already expressed at high levels at baseline. Shahi et al. reported the absence of any significant relationship between efflux pump expression levels and MIC values (P > 0.05) [38], consistent with these results. Of the 35 clinical isolates with rpoB mutations in this study, the MIC values for RIF did not rise with greater levels of efflux pump gene overexpression.

In this analysis, the extent to which the efflux pump inhibitor VP impacted MIC values for RIF was found to be related to the specific rpoB mutation sites. Following VP exposure, lower MIC reductions were observed in Mtb isolates harboring codon 445 mutations as compared to those harboring codon 450 or 170 mutations (Fig. 4). Higher levels of RIF resistance were thus maintained by these isolates with codon 445 mutations even following VP exposure. This may be because amino acid 445 contributes more significantly to RIF resistance such that strains with mutations affecting this residue may be able to preserve high levels of RIF resistance even following VP exposure. The 445 site mutations (His445Tyr, His445Asp, His445Arg) and 450 site mutations (Ser450Leu, Ser450Phe) in the RIF target gene rpoB have previously been confirmed to be related to high levels of RIF resistance [5, 39, 40]. However, no prior studies have demonstrated a greater contribution of codon 445 mutations to RIF resistance as compared to codon 450 mutations. While these data suggest the two may be closely related, the isolates included in this study did not exhibit other rpoB mutations with the potential to impact RIF resistance, including the notable lack of isolates with a point mutation at site 430 that may exhibit low levels of RIF resistance. Additional research is thus warranted to better elucidate this research and to clarify whether efflux pumps can play a role in the clinical treatment of tuberculosis.

In addition, the present study revealed a close relationship between exposure to the efflux pump inhibitor VP, the number of overexpressed efflux pump genes, and the reduction in the MIC values for RIF for the analyzed clinical Mtb isolates. Of the 35 RIF-resistant isolates included in this study, those with the greatest declines in MIC values tended to exhibit higher numbers of overexpressed efflux pump genes. This may be a consequence of the combined effects of efflux activity for multiple efflux pumps acting on RIF. However, no correlation between MIC values following VP treatment and the number of overexpressed efflux pump genes was detected. For example, in clinical isolates harboring the rpoB Ser450Leu mutation with a MIC of 16 μg/mL following VP treatment, the number of overexpressed efflux pump genes ranged anywhere from two to nine. No prior studies have reported on this point, emphasizing a need for further data in the future.

In total, the present study revealed that 11 efflux pump genes were downregulated more than fourfold in response to the combined pressure of RIF and VP, but no isolates exhibited > fourfold downregulation of Rv0451c (mmpS4) or Rv0450c (mmpL4), with just 17.1% of the isolates presenting with > twofold downregulation. Caleffi-Ferracioli et al. previously found that all of the efflux pump genes which they analyzed, including Rv1819c, Rv1258c, and Rv1410c, were downregulated following culture in the presence of VP and RIF for 72 h [41], in line with the present results. They also observed higher levels of Rv1819c (bacA), Rv1258c, and Rv2459 following culture for 16 h in the presence of RIF and VP relative to expression levels observed in response to RIF alone. In a separate study, Caleffi-Ferracioli et al. determined that VP had a better effect on the downregulation of efflux pump gene expression at 72 h relative to 16 h [20]. This may be a consequence of stress responses engaged by Mtb in response to short-term VP treatment, leading to the transient upregulation of efflux pump genes in an effort to eject this inhibitory compound from the cell. Over longer periods of time, however, gradual increases in the efficacy of VP culminate in efflux pump gene downregulation. Canezin et al. reported that Mtb cultured in the presence of RIF and VP for 72 h presented with significantly higher levels of Rv1819c (bacA) expression as compared to cells cultured in the presence of RIF alone, potentially owing to the risk of lethality to Mtb cells which may have driven the upregulation of this gene in an effort to facilitate survival via the excretion of RIF [42].

In summary, the present analysis of efflux pump gene expression in RIF-resistant Mtb cells harboring rpoB gene mutations revealed that efflux pumps serve as important mediators of RIF resistance in addition to these mutations within the drug target-coding gene. Rather than serving as mutually exclusive mechanisms of resistance, these two processes appear to cooperatively facilitate overall resistance phenotypes. While these analyses included only a relatively limited number of clinical isolates with different rpoB mutation sites, the mutation sites most commonly observed in clinical settings were included herein and therefore have a certain universality and representativeness.