Tuberculosis (TB) and drug-resistant tuberculosis (DR-TB) are global challenges, and their prevention and control are being prioritized by the World Health Organization (WHO) (WHO, 2020a). The cure rates of drug-susceptible TB, rifampicin-resistant (RR)/multidrug-resistant tuberculosis TB (MDR-TB), and extensively drug-resistant (XDR) TB were 85, 57, and 39%, respectively (WHO, 2019a, 2020a). The higher rate of unfavorable treatment outcomes observed with DR-TB might be due to a lack of effective drugs. In addition, the current treatment regimens for DR-TB might cause severe side effects (Ginsberg and Spigelman, 2007; Jacobson et al., 2010). For better management of MDR-TB, a government-organized and hospital-based management program for cases, denoted the Taiwan MDR-TB Consortium (TMTC), was established in 2007 (Yu et al., 2015), and the treatment success rate of MDR-TB increased significantly from 61% in the pre-TMTC era to more than 82% in the TMTC era (Yu et al., 2015). Nevertheless, new drugs are still needed for the management of difficult DR-TB cases.

Bedaquiline (BDQ), a diarylquinoline, is a novel antimycobacterial drug that was approved by the United States Food and Drug Administration (FDA) in 2012. BDQ was recommended by the WHO as a core drug for the treatment of MDR and XDR-TB in 2013 (Mahajan, 2013; WHO, 2013) and is part of the WHO-endorsed, shorter, and all-oral MDR-TB regimen (WHO, 2019b, 2020b). Since BDQ was classified as the priority drug (Group A) by the WHO for the treatment of MDR-TB in 2019, 109 countries have started using BDQ to treat MDR or XDR-TB by the end of the year (WHO, 2020a). BDQ shows efficiency with improved culture conversions in DR-TB treatment (Andries et al., 2005; Diacon et al., 2014; Nguyen et al., 2016). Additionally, BDQ exhibits no cross-resistance to current first-line and second-line anti-TB drugs except clofazimine (CFZ; Andries et al., 2005). However, since the introduction of BDQ for DR-TB treatment, BDQ-resistant TB strains have gradually emerged (Andries et al., 2014; Somoskovi et al., 2015; Veziris et al., 2017). Hence, it is necessary to adopt proper drug susceptibility testing (DST) for the prescription of prompt and adequate treatment.

Studies have revealed that the mechanisms that confer BDQ resistance to Mycobacterium tuberculosis mainly involves three genes, namely, the atpE (Andries et al., 2005), mmpR (Rv0678; Hartkoorn et al., 2014; WHO, 2021), and pepQ genes (Almeida et al., 2016). BDQ inhibits mycobacterial ATP synthase by targeting subunit C, which is encoded by the atpE gene, and the AtpE protein sequence is highly conserved (Andries et al., 2005). The gene variants A63P and I66M obtained from in vitro-selected mutants are associated with BDQ resistance (Andries et al., 2005; Petrella et al., 2006). Isolates harboring mutations in the atpE gene exhibit a relatively high minimum inhibitory concentration (MIC) to BDQ (10- to 128-fold; Nguyen et al., 2018). Mutations in the atpE gene cause failure in the binding of BDQ to subunit C of ATP synthase and thereby maintain the transfer of hydrogen ions and ATP production (Koul et al., 2007). In addition, the transcriptional repressor of the MmpS5-MmpL5 drug export pump is encoded by the Rv0678 gene (Andries et al., 2014). Mutations in Rv0678 cause upregulation of MmpS5-MmpL5 expression and the export of BDQ (Andries et al., 2014). Nevertheless, Rv0678 mutations are associated with low-level cross-resistance between BDQ and CFZ (Somoskovi et al., 2015) and lead to 2- to 8-fold increases in the MICs of BDQ and CFZ (Andries et al., 2014). Notably, a previous study highlighted that resistance to azole antifungal drugs is associated with Rv0678 mutations causing upregulation of the MmpS5-MmpL5 efflux pump (Milano et al., 2009). In addition, mutations in the pepQ gene, which encodes aminopeptidase, are associated with low-level BDQ and CFZ resistance (Almeida et al., 2016).

BDQ was introduced in Taiwan in 2014 for the treatment of DR-TB. Due to the alarming emergence of BDQ resistance, we established an algorithm for detecting BDQ resistance in our programmatic management of drug-resistant TB (PMDT) programs. In this study, we performed BDQ susceptibility testing and depicted the extent of BDQ resistance in DR-TB cases.

BDQ is a core drug in the treatment of DR-TB responsible for reducing mortality and improving outcomes (Olayanju et al., 2018; Schnippel et al., 2018; Mbuagbaw et al., 2019). The emergence of BDQ resistance raises concerns in the DR-TB control program. In this retrospective population-based study, the rate of BDQ resistance (MIC≥0.25μg/ml) among DR-TB cases without BDQ and CFZ exposure was found to be 3.1% (28/898) in Taiwan, whereas other studies found values of 1.0% in France (Veziris et al., 2017), 1.3% in Russia (Peretokina et al., 2020), 2.2–3.9% in China (Pang et al., 2017; Liu et al., 2020; Yang et al., 2020), and 2.3% in a multicountry population (Diacon et al., 2014; Pym et al., 2016; Villellas et al., 2017). Furthermore, we identified 77.3% (17/22) of MGIT-BDQ-resistant isolates harboring Rv0678 mutations (Supplementary Table S2), which might not be associated with CFZ cross-resistance (Xu et al., 2017; Ghodousi et al., 2019; Beckert et al., 2020), whereas the other studies found corresponding values of 50.0–66.7% in China (Pang et al., 2017; Yang et al., 2020), 66.7% in Australia (Martinez et al., 2018), 71.4% in Germany (Andres et al., 2020), 75.0% in France (Veziris et al., 2017), 100% in South Africa (Nimmo et al., 2020b), 100% in Russia (Zimenkov et al., 2017; Peretokina et al., 2020), and 100% in a multicountry population (Villellas et al., 2017). BDQ resistance might naturally occurred or during treatment with other anti-TB drugs (Yang et al., 2020) or previous use of antifungal drugs (Milano et al., 2009; Hartkoorn et al., 2014). In Taiwan, the 4% prevalence rate of azole-resistant Aspergillus fumigatus clinical isolates mainly emerged from the environment and during antifungal treatment (Wu et al., 2020), and its influence on BDQ resistance remains elusive. The reason for preexisting BDQ resistance is unknown and challenges the future use of BDQ in DR-TB treatment.

Notably, BDQ resistance determined using by MGIT DST using the suggested that a≥1day cutoff value might yield disputable susceptibility results. We observed that 22.7% (5/22) of MGIT-BDQ-resistant isolates had no mutations in the atpE, Rv0678, pepQ genes, and the same observations were obtained in France, China, and Iran (Pang et al., 2017; Veziris et al., 2017; Ghajavand et al., 2019; Liu et al., 2020). Due to the identification of 5 MGIT-BDQ broadline-resistant isolates with genotypic WT results for Rv0678, the assessment of raw MGIT-BDQ DST data is suggested. Nevertheless, MGIT-BDQ-resistant isolates with no mutations in the atpE, Rv0678, and pepQ genes might be caused by other resistance mechanisms, such as non-Rv0678 transcriptional regulators of mmpL5/mmpS5, as proven in a system consisting of two components, TrcR and TrcS, using whole-genome microarray technology (Wernisch et al., 2003), or overexpression of the BDQ-response regulons Rv0324 and Rv0880 (Peterson et al., 2016).

Excluding the five aforementioned isolates, the susceptibility to BDQ determined using the WHO interim critical concentration for MGIT (1μg/ml) was reliable (WHO, 2018). Nevertheless, we found that an MIC of 0.5μg/ml, but not 0.25μg/ml, could consistently determine BDQ resistance based on mutations in Rv0678 (Table 3). Consequently, an MIC of 0.25μg/ml could be considered an intermediate-susceptible category for DST as defined by the Clinical Laboratory Standards Institute.

Previous studies have revealed that mutations scattered across Rv0678 result in MIC shifts and might not be linked to specific M. tuberculosis lineages (Villellas et al., 2017; Zimenkov et al., 2017; Ismail et al., 2019; Battaglia et al., 2020; Peretokina et al., 2020; Nimmo et al., 2020b). In this study, Rv0678 mutations were not associated with specific genotypes. Of note, we identified four Beijing isolates with whole Rv0678 gene deletion that exhibited various MICs ranging from 0.012 to 0.5μg/ml, and whether this deletion is a lost-of-function mutation or due to an existing epistatic factor merits further investigation. The intergenic region mutation c-11a (MIC=0.015μg/ml), which is found exclusively in Beijing isolates, was consistent with that observed in a study conducted in Belgium; however, it might not be associated with drug resistance (Villellas et al., 2017). Because isolates harboring the G87R mutation in Rv0678 are susceptible to BDQ (Martinez et al., 2018; Battaglia et al., 2020), a novel G87A mutation (MIC=0.03μg/ml) identified in this study might not have impacted the structure and stability of the protein. Studies have revealed that mutations occurring at amino acids 62–68 interfere with helix recognition of the DNA-binding domain in other MarR family regulators (Hong et al., 2005). We found that in 2 BDQ-resistant isolates, the introduction of Del 11–63/Fs (29 stops; N=1) and Ins g 181–182/Fs (80 stops) might cause loss of the functional folded protein and subsequently destabilize Rv0678 (Kadura et al., 2020). The association of drug resistance and novel Rv0678 mutations found in MGIT-BDQ-resistant isolates merits further investigation.

Furthermore, resistance-conferring mutations in the atpE gene and other probable BDQ resistance-associated genes, pepQ, Rv1979c (Ismail et al., 2018), and mmpL5, might be potential determinants. Because atpE and pepQ mutations were found to not confer high- or low-level BDQ resistance in this study and the existing Rv0678 mutations might not be associated with BDQ resistance, careful evaluation of the prescription of BDQ in the regimens for DR-TB treatment is recommended. In our PMDT program, periodical expert consultation on treatment and management is conducted through the TMTC.

Because most Rv0678 mutations are associated with low-level BDQ resistance (Veziris et al., 2017; Villellas et al., 2017; Xu et al., 2017; Zimenkov et al., 2017; Battaglia et al., 2020; Peretokina et al., 2020; Yang et al., 2020), scarce studies have investigated their impact on treatment outcomes. A study using a mice model showed that BDQ still exhibits bactericidal activity against isolates with Rv0678 mutations and activity lower than that found in the absence of Rv0678 mutations (Andries et al., 2014). Of the five cases with Rv0678 mutations that have not been exposed to BDQ and CFZ in South Africa, two cases exhibited favorable outcomes after treatment with BDQ-containing regimens (Nimmo et al., 2020a). Of the five patients who acquired BDQ resistance with Rv0678 mutations after BDQ treatment, four had unfavorable outcomes (Nimmo et al., 2020a). A study conducted in China showed that two BDQ- or CFZ treatment-naïve cases with Rv0678 mutations showed favorable outcomes after BDQ treatment (Liu et al., 2020). Nevertheless, of the five cases that acquired BDQ resistance with Rv0678 mutations after BDQ treatment, three cases had unfavorable outcomes (Liu et al., 2020). This was a retrospective cohort study, and we report observed results. Individualized regimens for 22 MGIT-BDQ-resistant TB cases were in Supplementary Table S3. Nevertheless, in line with WHO recommendations, we observed that DR-TB cases with isolates harbored Rv0678 mutations could be treated with nonBDQ-containing regimens of at least four drugs, and had favorable outcomes (Supplementary Table S3). Since the frequencies of favorable outcomes in other studies with limited numbers of TB cases, our collective results provide insights for DR-TB management. Particularly, of the five primary BDQ-resistant cases treated with a regime that included BDQ, four cases had favorable outcomes, and one was under treatment with sputum culture conversion (Supplementary Table S2). No case of relapse was recorded after 7years of BDQ use.

This study provides the first report on the population-based surveillance and molecular characteristics of BDQ susceptibility among DR M. tuberculosis isolates in Taiwan. For isolates with borderline phenotypic resistance to MGIT-BDQ, checking for GU differences or conducting genotypic analyses are suggested to rule out the possibility of BDQ resistance. In accordance with other previous studies, the present study found that BDQ resistance was mainly caused by Rv0678 mutations, including some that were not resistance-conferring mutations. Furthermore, a MIC of 0.25μg/ml could be considered an intermediate-susceptible category for DST. We observed favorable outcomes among patients with DR-TB receiving BDQ-containing regimens regardless of Rv0678 mutations. In the PMDT program, comprehensive DST should be performed to inform the prescription of BDQ in the treatment of DR-TB with the aim of achieving better treatment outcomes.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

RJ designed the research. S-HW, C-HC, and H-CH performed the experiments. RJ and S-HW analyzed the results and wrote the manuscript. All authors contributed to the article and approved the submitted version.

This study was supported by grants (MOHW109-CDC-C-315-114403 and MOHW110-CDC-C-315-114405) from the Taiwan Centers for Disease Control, Ministry of Health and Welfare, Taiwan.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.