Introduction

Tuberculosis (TB), an infection caused by Mycobacterium tuberculosis, remains one of the deadliest infectious diseases, responsible for 1.25 million deaths in 2023[1](https://www.nature.com/articles/s41467-025-64427-6#ref-CR1 “World Health Organisation. Global Tuberculosis report 2024. https://www.who.int/teams/global-tuberculosis-programme/tb-reports/global-tuberculosis-report-2023

(2024).“). The severity of the global TB pandemic is further exacerbated by the emergence of multidrug-resistant TB (MDR-TB). Effective treatment of MDR-TB typically requires a combination of three to five drugs, many of which are associated with toxicity and severe adverse events. These side effects often lead to poor adherence to treatment. Combination treatment is essential to shorten treatment duration, minimise the risk of drug resistance and prevent relapse. Nevertheless, current regimens still require many months of treatment with prolonged therapy contributing to reduced patient compliance and an increased risk of drug resistance.

Considerable research efforts are focused on optimising drug combinations that enable treatment shortening. For example, the TB-PRACTECAL clinical trial aimed to identify combination regimens, given as a 6 month oral regimen, capable of shortening treatment for DR-TB2,3. The final results demonstrated that the BPaLM (bedaquiline, pretomanid, linezolid and moxifloxacin) regimen was both safer and more effective, achieving more favourable outcomes (88%) in patients compared with the standard of care (SoC; 59%). Similarly, BPaL (86%) and BPaLC (bedaquiline, pretomanid, linezolid, clofazimine; 77%) regimens also resulted in higher proportions of MDR-TB patients with favourable outcomes, supporting the potential of shorter, optimised therapies using these compounds.

The World Health Organisation’s consolidated 2024 guidelines now include new recommendations for MDR-TB treatment shortening[4](https://www.nature.com/articles/s41467-025-64427-6#ref-CR4 “World Health Organisation. Who Consolidated Guidelines On Tuberculosis. Module 4: Treatment - Drug-resistant Tuberculosis Treatment. https://www.who.int/publications/i/item/9789240063129

(2022).“),[5](https://www.nature.com/articles/s41467-025-64427-6#ref-CR5 “World Health Organisation. Key Updates To The Treatment Of Drug-resistant Tuberculosis - Rapid Communication. https://www.who.int/publications/i/item/WHO-UCN-TB-2022-2

(2024).“). These recommendations suggest two 6 month regimen options (compared with the previous 9 to 18 month duration): BPaLM for ages 14 years and older, and BDLLfxC (bedaquiline, delamanid, linezolid, levofloxacin and clofazimine) for children, adolescents, adults and pregnant women. However, the central role that fluoroquinolones and linezolid play in those regimens poses serious limitations due to drug resistance and toxicity, respectively. Moxifloxacin or levofloxacin must be excluded in cases of fluoroquinolone resistance, which is rapidly becoming a major challenge for treating MDR-TB6. This results in the modification of these regimens to BPaL and BDLC, respectively. Furthermore, the use of linezolid in MDR-TB treatment is often associated with adverse events, such as myelosuppression and peripheral neuropathy, which may negatively impact patient adherence. Thus, although MDR-TB treatment options are diversifying and improving[7](https://www.nature.com/articles/s41467-025-64427-6#ref-CR7 “Stop T. B. Partnership. Clinical Pipeline. https://www.newtbdrugs.org/pipeline/clinical

(2024).“), there is growing interest in identifying alternatives to the fluoroquinolones and linezolid to partner with BPa8,9,10,11. Furthermore, emerging drug resistance adds additional concerns, highlighting the urgent need to discover compounds with novel modes of action (MoA) and develop innovative therapeutic approaches.

The discovery of telacebec (Q203) and its successful Phase 2 A Early Bactericidal Activity (EBA) trial has established inhibition of the M. tuberculosis cytochrome bc1 complex as a promising new drug target12,13,14,15. As a crucial component of the electron transport chain, essential for ATP production, cytochrome bc1 represents an attractive partner target for other inhibitors of oxidative phosphorylation, such as bedaquiline. However, cytochrome bd, a terminal oxidase predominantly active in low-oxygen environments, has been identified as a critical bypass mechanism that can sustain bacterial respiration under cytochrome bc1 inhibition16. This has led to growing interest in dual inhibition strategies and positioned cytochrome bd as another underexploited target in TB drug development. While cytochrome bc1 inhibitors have demonstrated considerable potential, their specific role within future TB treatment regimens is not clearly defined17.

While counting colony-forming units (CFU) can provide a measure of bacterial burden during treatment, it does not capture the ability of a regimen to fully sterilise infection and prevent relapse. A relapse study, which assesses bacterial regrowth after treatment cessation, provides a more stringent and clinically relevant measure of treatment efficacy. This is particularly important in TB, where surviving M. tuberculosis bacilli may enter a non-replicating or persistent state and later cause disease recurrence. Therefore, relapse models offer critical insights into the sterilising activity of drug combinations, which is a key consideration for regimen design aimed at treatment shortening.

Here, using relapsing TB mouse models, we assessed the contribution of validated cytochrome bc1 inhibitors, including the tool compound JNJ-290118, in different treatment regimen strategies to assess their potential to replace and mitigate the liabilities of the fluoroquinolones and linezolid. Next, we demonstrate that a cytochrome bc1 inhibitor could contribute to an ultrashort treatment for drug sensitive TB (DS-TB), potentially contributing to a ≤2 month treatment regimen. Finally, we demonstrate that cytochrome bc1 inhibitors are significantly more bactericidal against M. tuberculosis clinical isolates compared to laboratory-adapted strains, suggesting future drug discovery efforts should consider more emphasis on the testing of clinical isolates during drug development. Our findings suggest that cytochrome bc1 inhibitors could provide an important contribution to TB treatment regimens by enhancing sterilisation, reducing relapse rates, and improving treatment outcomes.

Results

Cytochrome bc 1 inhibitors as alternative partner drugs in MDR-TB treatment regimens

Drug resistance to fluoroquinolones, along with the adverse events associated with linezolid, has prompted the search for alternative partner drugs for MDR-TB treatment. In this context, we recently identified JNJ-2901, an analogue of telacebec, as a promising candidate for further development of best-in-class inhibitors targeting cytochrome bc1 (Supplementary Table S1–3). JNJ-2901 exhibited sub-nanomolar activity against MDR-TB strains and achieved a 4-log reduction in bacterial burden using a M. tuberculosis cytochrome bd knockout (CytBd-KO) strain in an acute mouse model of TB infection18. Cryo-electron microscopy structural analysis revealed that JNJ-2901 occupies the same binding pocket as telacebec and other structurally related cytochrome bc1 inhibitors, confirming a shared MoA. All inhibitors described in this study were evaluated for their absorption, distribution, metabolism and excretion (ADME) properties, pharmacokinetics (PK) and toxicology profiles, supporting their further use in subsequent experiments14 (Supplementary Table S2–3).

Using a relapsing mouse model, we compared the efficacy of cytochrome bc**1 inhibitors, including JNJ-2901 and telacebec, in various combination regimens (Fig. 1a, Supplementary Fig. S1). We first focused on replacing linezolid with JNJ-2901 (J) in the presence and absence of either moxifloxacin (M) or clofazimine (C) in the BPaL regimen, the recommended SoC for fluoroquinolone-resistant MDR-TB. Only BPaMZ provided a completely sterile cure after 8 weeks treatment (defined as 0/15 mice; 0%), measured after 12 weeks of relapse post treatment. However, BPaCJ (33%) was superior to BPaL (87%) based on relapse rates. When combined with JNJ-2901, the addition of clofazimine (BPaCJ; 33%) resulted in a statistically significant reduction in relapse rates compared to the addition of moxifloxacin (BPaMJ; 100%; p = 0.003) (Fig. 1b; Supplementary Table S4–5; Study A).

Fig. 1: Investigating alternative MDR-TB treatment regimens based on inclusion of a cytochrome bc1 inhibitor in relapsing mouse models of TB.

a Schematic of the studies presented in this project. b–d Lung bacterial burden, proportions of mice with positive cultures at end of treatment and relapse in M. tuberculosis-infected mice from Studies A, B and D. CFU data is shown for week 8 (b, c) and week 12 (d). Details of treatment doses can be found in Supplementary Tables S4 and S11. bedaquiline: B; pretomanid: Pa; clofazimine: C; linezolid: L; moxifoxacin: M; pyrazinamide: Z; JNJ-2901: J. SoT: Start of treatment. *: relapse rate; 12 or 16 weeks after treatment cessation ( + 12 or 16 wks). Relapse rate (n/N). Source data are provided as a Source Data file.

In a second relapse study, we focused on the impact of clofazimine in more detail (Fig. 1c; Supplementary Table S4, S6–7; Study B). BPaL led to a greater decrease in CFU after 8 weeks compared to BPaJ, but with comparable relapse rates observed after 8 weeks (100% for BPaL and BPaJ) and 12 weeks of treatment (27% for both regimens; Fig. 1c; Supplementary Table S4, S6–7). Replacing linezolid with clofazimine resulted in a similar decrease in CFU after 8 weeks but led to fewer relapses post-treatment (100% for BPaL, 27% for BPaC). BPaCJ demonstrated the best bactericidal effect, achieving an additional 1 log10 decrease in CFU compared to BPaC after 8 weeks, translating to a 33% relapse rate. No relapses occurred after 12 weeks of treatment with either BPaC or BPaCJ, highlighting the potential for treatment shortening compared to the BPaL SoC (Fig. 1b, c; Supplementary Table S4–7). These findings demonstrate that combining a cytochrome bc1 inhibitor with BPaC increases the sterilising activity compared to BPaC alone, suggesting that cytochrome bc1 inhibitors could play an important role in future MDR-TB regimens.

In an additional study (Supplementary Table S8–10; Study C), JNJ-2901 reduced the relapse rate when combined with the BPaL regimen, despite an inability to rescue mice when administered as a monotherapy in an intravenous model using the H37Rv reference strain (Supplementary Fig. S2). There was a trend towards improvement of both the bacterial burden (BPaLJ: 0.95 ± 1.1 versus BPaL: 2.06 ± 1.3 log10 CFU lung−1, p = 0.04) and a quantifiable improvement in relapse rates, that failed to reach statistical significance within the limitations of this trial design (50% versus 89%, p > 0.05). Notably, BPaLJ was comparable to BPaLM and BPaMZ (Supplementary Table S8–10). We recommend future studies include increased replicates to provide greater statistical power. Nevertheless, this suggests that BPaLJ could be considered as a potential treatment regimen irrespective of fluoroquinolone susceptibility.

Sterilisation potential of cytochrome bc 1 inhibitors in an alternative dosing regimen

We further evaluated the role of cytochrome bc1 inhibitors in an alternative dosing strategy, where mice were treated with BPaL for 8 weeks during an initial phase, followed by an additional 8 weeks of either BPa, BJ, B or no treatment in a continuation phase (Fig. 1d; Supplementary Table S11; Study D). The rationale for this study design was to reduce the time and overall amount of drug required to reach bacterial sterility. Bacterial burdens in lungs were assessed after 8, 12 and 16 weeks, and relapse rates were measured after 16 weeks of treatment plus 16 weeks following treatment cessation. Both the BPaL/BPa and BPaL/BJ regimens trended towards lower relapse rates compared to BPaL without continuation treatment, although this was not statistically significant (p = 0.09). No colonies were detected after 12 weeks of treatment with BPaL/BJ, whereas with BPaL/B, colonies were detected from 3/5 mice at the end of treatment (Fig. 1d; Supplementary Table S11). These results suggest that inclusion of a cytochrome bc1 inhibitor could enhance regimen effectiveness during the continuation phase of treatment with fewer drugs and a reduced overall drug burden.

Inclusion of cytochrome bc 1 inhibitors can lead to treatment shortening

Next, we focused on an ultra-short treatment strategy for DS-TB based on drugs targeting the respiratory pathway. In the recent TRUNCATE-TB trial, a 2 month BZ-containing regimen was shown to be as effective as the current 6 month SoC (isoniazid, rifampicin, pyrazinamide, and ethambutol; HRZE)19. Furthermore, we previously demonstrated that an in vitro combination of bedaquiline, clofazimine and telacebec (T), which all target components of the electron transfer chain, resulted in rapid killing, suggesting a strong basis for an ultra-short treatment20. To further investigate this in vivo, we assessed the efficacy of regimens containing telacebec alongside other drugs targeting the respiratory pathway, compared to the current DS-TB SoC (HRZE; Study E). BCZ, CZT and BCZT regimens all demonstrated superior efficacy over HRZE after 8 weeks of treatment in reducing lung CFU burdens (Fig. 2, Supplementary Tables S12–13; Supplementary Data 1; Study E). Both BCZ and BCZT regimens achieved 0% relapse rates 12 weeks after 6- and 8 weeks of treatment, whereas CZT required at least 12 weeks of treatment to achieve 0% relapse. One of the mice treated with HRZE remained culture positive even after 20 weeks of treatment. The benefit of adding telacebec to BCZ is shown by the relapse rates after 4 weeks treatment: BCZ (60%) versus BCZT (36%), highlighting the treatment-shortening potential of cytochrome bc1 inhibitors in combination with other drugs targeting the electron transport chain. Adding telacebec to the CZ core regimen reduced bacterial load by 2.2 log10 CFU and dramatically improved relapse rates after 12 weeks treatment from 100% (CZ) to 0% (CZT; p < 0.0001). Furthermore, CZT, BCZT and BCZ regimens significantly shortened treatment compared to HRZE (Fig. 2; Supplementary Tables S12–13; Supplementary Data 1). These findings provide strong evidence that incorporating a cytochrome bc1 inhibitor into new treatment regimens offers substantial benefits which should be further investigated.

Fig. 2: Investigating alternative DS-TB treatment regimens based on inclusion of a cytochrome bc1 inhibitor.

Lung bacterial burden and relapse in M. tuberculosis-infected mice from Study E. Details of treatment doses can be found in Supplementary Table S12. Limit of detection (dotted line) was 0.4 log10 CFU lung-1. Bedaquiline: B; clofazimine: C; rifampicin: R; ethambutol: E; isoniazid: H; telacebec: T; pyrazinamide: Z. Source data are provided as a Source Data file.

Increased susceptibility of cytochrome bc 1 inhibitors against clinical isolates

We found that the impact of cytochrome bc1 inhibitors may be underrepresented depending on the M. tuberculosis strain selected for investigation. Here, using a range of cytochrome bc1 inhibitors, including JNJ-4052, which has similar activity, ADME and PK profiles as JNJ-2901 (Supplementary Fig. S1; Supplementary Tables S1–3), we demonstrate that TB clinical isolates exhibited increased susceptibility to cytochrome bc1 inhibition. This was observed against a diverse panel of clinical isolates from infected patients in both minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) assays, when compared to the laboratory-adapted H37Rv strain (Fig. 3a–d; Supplementary Table S14). This heightened susceptibility translated into increased in vivo efficacy in an acute mouse model, where a statistically significant CFU reduction (1.6 log10 reduction; p < 0.01) was observed following cytochrome bc1 inhibitor (JNJ-4052) treatment in mice inoculated with a clinical isolate (Fig. 3e; Supplementary Tables S1–S3; Study F). These results further emphasise the importance of including clinical isolates when evaluating the in vivo efficacy of new compounds.

Fig. 3: Bactericidal activity of cytochrome bc1 inhibitors in clinical isolates.

a Distribution of MIC90 (concentration achieving 90% inhibition) values for a diverse selection of clinical isolates (black) compared with the lab-adapted WT H37Rv (cyan) for JNJ-4052 and bedaquiline (BDQ) (Supplementary Table S14). Clinical isolate N1283 (orange) is indicated. n = 2 technical replicates. Impact of (b) JNJ-4052, (c) JNJ-2901 and (d) telacebec on CFU counts in WT H37Rv compared to N1283 clinical isolate. n = 4 biological replicates for (b−d). e In vivo efficacy (Study F) of JNJ-4052 (50 mg kg-1; PO) in H37Rv and N1283 after 2 weeks of treatment. SoT: Start of treatment, 7 days after inoculum with 300 CFU. n = 3 mice. Significance calculated with an ordinary one-way ANOVA. Data are presented as geometric mean ± geometric SD (panels b−d) or mean ± SD (e). Source data are provided as a Source Data file.

Discussion

Our findings highlight the potential of cytochrome bc1 inhibitors as valuable components of future TB treatment regimens. These compounds may serve as effective alternatives to linezolid in SoC regimens for MDR-TB and fluoroquinolone-resistant MDR-TB. Given their novel MoA, cytochrome bc1 inhibitors may also contribute to raising evolutionary barriers to resistance when used in combination and may be suitable for inclusion in DS-TB regimens to reduce treatment duration.

Substitution of linezolid with JNJ-2901 within BPaL regimens demonstrated promising results. While no regimen achieved complete sterilisation after 8 weeks of treatment, the BPaCJ combination significantly reduced relapse rates compared to BPaL and BPaMJ, demonstrating the additive benefit of clofazimine over moxifloxacin when paired with a cytochrome bc1 inhibitor. Replacing linezolid with clofazimine (BPaC) also led to improved post-treatment outcomes. The further addition of a cytochrome bc1 inhibitor (BPaCJ) not only reduced bacterial burden more effectively than BPaC alone but also achieved relapse-free cure after 12 weeks, indicating enhanced sterilising activity and the potential for treatment shortening. Collectively, our findings support further evaluation of cytochrome bc1 inhibitors as promising components of future MDR-TB regimens.

Effective TB treatment regimens require a combination of drugs with both bactericidal and sterilising activities to rapidly reduce bacterial loads and eliminate drug-tolerant bacilli. A two-phase regimen, with an initial intensive phase of drugs that rapidly reduce bacterial burdens, followed by sterilising drugs in a continuation phase, is a proven strategy for improving treatment outcomes for patients. Drugs in the continuation phase must target hard-to-treat drug-tolerant bacteria that survive the initial intensive phase21. In our study, the inclusion of a cytochrome bc1 inhibitor in the continuation phase of the BPaL/BJ regimen demonstrated a sterilising effect similar to that of the BPaL/BPa regimen, suggesting that a cytochrome bc1 inhibitor could serve as an effective alternative to pretomanid. Furthermore, a two-phase regimen could facilitate the development of long-acting injectable formulations, offering an alternative strategy to improve treatment adherence and outcomes[22](https://www.nature.com/articles/s41467-025-64427-6#ref-CR22 “EU Clinical Trial (2023-508810-41-00). A Single Ascending Dose, Single-Centre Study, to Assess Pharmacokinetics, Safety and Tolerability of a Single Intramuscular Dose of Bedaquiline Long-Acting Injection Formulation in Healthy Participants. https://euclinicaltrials.eu/ctis-public/view/2023-508810-41-00

(2024).“).

There is an urgent need to develop ultra-short TB treatment regimens for drug-sensitive TB to improve patient compliance and reduce the risk of drug resistance. Bedaquiline-containing regimens have emerged as the most promising approach to achieve this goal. Additionally, a 14 day Phase IIA clinical trial previously demonstrated that pyrazinamide enhances the efficacy of bedaquiline-containing regimens, thereby contributing to treatment shortening23. Here, we show that the inclusion of a cytochrome bc1 inhibitor dramatically reduced treatment duration, particularly in comparison to the current SoC (HRZE). Furthermore, a BCZ-background, especially when enhanced with telacebec, showed potential to achieve treatment durations of under 2 months. Further evaluation in larger preclinical and clinical studies is required to confirm long-term efficacy and relapse rates.

An interesting finding from our study was the increased bactericidal activity of cytochrome bc1 inhibitors against clinical isolates and a cytochrome bd knockout strain. Similar results have recently been reported, showing that the clinical isolate M. tuberculosis HN878 is more susceptible to cytochrome bc1 inhibitors24. This study demonstrated that the addition of telacebec to BPaC in HN878 led to a 2-log reduction in lung bacterial burden compared to BPaC alone after 4 weeks of treatment. This likely reflects differences in the expression of cytochrome bd after cytochrome bc1 inhibition. Clinical isolates are reported to have a lower basal level of expression of cytochrome bd than the lab-adapted strain H37Rv25. Under low bd expression or in knockout conditions, the bacterium is forced to rely more heavily on cytochrome bc1 for respiration, making it more vulnerable to inhibition16. This is evident from the MBC99.9 data for H37Rv, where cytochrome bc1 inhibitors are primarily bacteriostatic, whereas in the H37Rv_CytBd-KO strain, the same inhibitors are bactericidal (Supplementary Table S2). Taken together, these findings suggest the choice of strain can influence the apparent drug activity and support the routine inclusion of clinical isolates in early drug discovery and regimen design studies.

While our study provides important insights into the activity of novel combinations containing cytochrome bc1 inhibitors, there are several considerations that can inform future experimental design. First, some experiments were underpowered to detect statistically significant differences between conditions. Future studies will increase animal numbers to strengthen statistical confidence and maximise interpretability. Second, we did not include monotherapy arms for each compound in the combinations tested, which limits our ability to deconvolute the contribution of individual agents. Including monotherapy controls in future work will provide a clearer understanding of synergistic or additive effects. Third, only female mice were used in this study to simplify animal husbandry and reduce stress-related variables as male mice require separation and can experience negative effects from isolation, increasing welfare concerns and experimental attrition. While sex-based differences have been reported in clinical outcomes26, and some sex-specific immunological differences have been reported for different infectious diseases27,28, gender bias remains unclear when studying the therapeutic effect of anti-tubercular drugs. As our goal was to provide proof-of-concept data rather than clinical validation, we chose a single-sex design for consistency, reproducibility and humane conduct of our multi-month studies; however, future studies should explicitly investigate potential sex differences, particularly when advancing regimens toward clinical development. Finally, our findings highlight that clinical isolates can exhibit markedly different susceptibility profiles compared to the widely used H37Rv lab-adapted strain. Future work should consider including diverse clinical strains to better capture the variability seen in patient-derived M. tuberculosis populations and improve translational relevance.Building on the established early bactericidal activity and novel mechanism of action of telacebec, cytochrome bc1 inhibitors continue to show promise as key components in future TB treatment regimens. Their ability to enhance sterilising activity, particularly in combination with bedaquiline and clofazimine, supports their potential role in ultra-short TB treatment regimens. Beyond TB, telacebec is also under investigation for the treatment of Buruli ulcer and leprosy[29](#ref-CR29 “TB Alliance. New Clinical Trial Examines Use Of Novel Compound Telacebec In Buruli Ulcer Patients. https://www.tballiance.org/news-new-clinical-trial-examines-use-novel-compound-telacebec-buruli-ulcer-patients/

(2024).“),30,31.

Cytochrome bc1 inhibitors not only improved the bactericidal and sterilising activity of regimens in mouse relapse models but also showed enhanced potency against clinical isolates. Our data highlight the potential of both clofazimine- and cytochrome bc1 inhibitor-containing regimens to significantly shorten treatment durations for both DS- and DR-TB. This work serves as a foundation for further exploration of other treatment regimens incorporating cytochrome bc1 inhibitors and there are several in vitro, ex vivo and in vivo tools available to expedite this process. For example, pairwise drug interactions can be rapidly assessed using DiaMOND analysis to identify optimal combinations in vitro32. These combinations can then be tested in ex vivo caseum models to investigate the contribution of cytochrome bc1 inhibition against non-replicating, persistent bacteria33. Promising regimens could subsequently be evaluated in an ultra-short-course treatment model that incorporates RS-ratio measurements to rapidly assess bactericidal activity and sterilising potential34.

These findings support the inclusion of cytochrome bc1 inhibitors in future regimen design efforts and highlight the need for further studies to define optimal dosing strategies, assess long-term relapse outcomes and evaluate their role in additional combinations.

Method

Treatment regimens were evaluated in five studies (Studies A–E), in a relapsing mouse infection model of TB treatment35. Animal studies were performed at Evotec France SAS Toulouse (Studies A and B), Sorbonne University (Study C)36, Johns Hopkins University (Study D) and Colorado State University (Study E), in accredited facilities. An additional study (Study F) in an acute mouse infection model was performed at London School of Hygiene & Tropical Medicine37,38.

Ethical statement

In Studies A and B, animal experiments were performed under the European Union Directive and with local ethical committee clearance. The study procedures were reviewed by the Evotec France Ethical Committee and authorised by the French Ministry of Education Advanced Studies and Research. For Study C, the experimental project was favourably evaluated by the ethics committee n°005 Charles Darwin localised at the Pitié-Salpêtrière Hospital and clearance was given by the French Ministry of Education and Research under the number APAFIS#12380-2017112809414820 v3. The animal facility received the authorisation to carry out animal experiments (license number C-75-13-08). For Study D, ethics oversight was provided by the Johns Hopkins University Animal Care and Use Committee, which is PHS assured, USDA registered, and AAALAC accredited. For Study E, ethics oversight was provided by the Colorado State University Animal Care and Use Programme (reference number, KP 5172) which is PHS assured, USDA registered, and AAALAC accredited. For Study F, animal procedures were performed under UK Home Office project license P6CA9EB8D and approved by the London School of Hygiene & Tropical Medicine Animal Welfare Ethical Review Board. All work was conducted in accordance with the UK Animal Scientific Procedure Act (ASPA) 1986. In all cases, the humane end points were: (i) body weight loss ( ≥ 20% weight loss), (ii) body condition scoring (behavioural changes), (iii) dehydration and (iv) laboured respiration (potentially due to exacerbated lung pathology).

Animal housing

Female BALB/cJRj mice (studies A, B, D, E, and F) and female Swiss mice (Study C) were housed socially in bio-confined BSL3 cages (up to 5 animals/cage) under a 12-h light: 12-h dark with access to filtered water and a standard rodent diet ad libitum. An ambient temperature of 22 ± 2 °C, a relative humidity of 55 ± 10% and a negative pressure of -20Pa were maintained. All mice were allowed to acclimatise to their new environment for at least 5 days after identification. Mice were observed daily.

Study design

In Studies A and B, mice were infected by intranasal inoculation of 50 µL of M. tuberculosis H37Rv at inoculum level of 4.5 log10 CFU mouse−1. Treatment was initiated 2 weeks post-infection, when the bacterial burden in the lungs was >7.19 log10 CFU. The mice were treated 5 days per week by oral gavage (at 10 mL kg−1), for 4, 8, and 12 weeks and received different treatment combinations (Supplementary Table S4) containing bedaquiline (25 mg kg−1), pretomanid (40 mg kg−1), linezolid (100 mg kg−1), clofazimine (20 mg kg−1), and JNJ-2901 (5 mg kg−1). For all groups, the first two drugs were administered in the morning and the last two drugs in the afternoon, with around 2 h between the administration of each drug. The bacterial load in lungs (CFU lung−1) was assessed at the end of each treatment, after a standard washout period of 3 days, in 5 animals/group. An untreated control group consisted of 10 mice (5 at D-13 to act as infection control and 5 at D0 to determine the infection level at treatment start). At the end of 8 or 12 weeks of treatment, 15 animals per group were held without treatment for 12 weeks to determine the proportion of mice with relapse. The total number of mice was 160 in Study A and 255 in Study B. Other study endpoints included animal weights weekly prior to treatment and post treatment during the relapse period and 3 times weekly during treatment phase, lung weight 1 day and 14 days pi and at the end of each period of treatment and relapse period, and observation of treatment-emergent adverse effects.

For Study C (Supplementary Table S8), the M. tuberculosis reference strain H37Rv maintained for animal infection was grown on Löwenstein-Jensen medium for 3 weeks. Colonies were subcultured in 7H9 medium supplemented with 10% OADC for 7 days at 37 °C. 112 six-week-old female Swiss mice were purchased from Janvier breeding centre. Mice were infected intravenously with 0.5 mL of the bacterial suspension of 6.1 log10 CFUs. Mice were treated orally 5 days per week for 3 months with the following drugs (dosing in mg/kg/day): bedaquiline (B, 25 mg kg−1), JNJ-2901 (J, 10 mg kg−1), moxifloxacin (M, 100 mg kg−1), pretomanid (Pa, 40 mg kg−1), linezolid (L, 100 mg kg−1), rifampicin (R, 10 mg kg−1), isoniazid (H, 25 mg kg−1) and pyrazinamide (Z, 150 mg kg−1). Six untreated mice were sacrificed the day after infection (D-13), the day of treatment initiation (D0); 20 mice were allocated in each arm and were euthanized after 16 weeks of treatment and 12 weeks off treatment for relapse rate assessment.

In Study D, 4–6 weeks old mice were infected by high-dose aerosol with M. tuberculosis H37Rv (~4 log10 CFU mL−1), using a Glas-Col Inhalation exposure system. Treatment started at 2 weeks pi (when the bacterial burden was >7 log10 CFU) and was administered by oral gavage, once daily, 5 days per week. Mice received different treatment combinations (two or three drugs; Supplementary Table S11) containing bedaquiline (25 mg kg−1), pretomanid (100 mg kg−1), linezolid (100 mg kg−1) and JNJ-2901 (5 mg kg−1). All treatment groups received BPaL for an initial 8-week intensive phase, then received treatment with B alone or in combination with either Pa or J, or no treatment, for an 8-week continuation phase. The untreated control group consisted of 26 mice (eight at D-13, eight at D0, and 5 at each 2- and 4-weeks after the start of treatment). Lung CFU was assessed after 8, 12 and 16 weeks of treatment (5 mice/group), as well as at the end of the relapse period (15 mice/group).

In Study E, 6–8 weeks old mice were infected by high-dose aerosol with M. tuberculosis Erdman, using a Glas-Col Inhalation exposure system. Treatment was initiated at 11 days pi and was administered 5 days per week, via oral gavage at 200 µL/mouse, ensuring at least 1 h between the administration of regimen components. Mice received different treatment combinations (2–4 drugs; Supplementary Table S12) containing bedaquiline (25 mg kg−1), clofazimine (20 mg kg−1), pyrazinamide (150 mg kg−1), telacebec (10 mg kg−1), isoniazid (10 mg kg−1), rifampicin (10 mg kg−1), and ethambutol (100 mg kg−1) for 4, 6, 8, 12, 16, and 20 weeks. The untreated control group consisted of 11 mice (six sacrificed at D-11 and 5 at the start of treatment on D0). Lung CFU was assessed at the end of the treatment period in 5 mice/group after a washout period of 5 days. Relapse groups were kept without treatment for 12 weeks; 15 mice/group were allocated for each arm. Mice were euthanised at the sampling timepoints indicated in Supplementary Table S12.

In Study F, 6–8-week-old BALB/c mice (Charles River UK) were infected by high-dose aerosol with 300 CFU of either M. tuberculosis H37Rv mouse passaged or a clinical isolate, N1283. After 7 days, treatment was initiated with once-daily JNJ-4052 (50 mg kg−1; PO) for 2 weeks. Lung CFU was assessed at the start of treatment (day 7) and after treatment (day 21) (3 mice/group).

Drug formulation

Bedaquiline (salt-base ratio 1.21; studies A and B: LTK Laboratories, B165121. Study C, D and E provided by Janssen Pharmaceutica) was prepared weekly. It was dissolved in 20% 2-hydroxypropyl-β-cyclodextrin (Aldrich 332593) at pH 3 (adjusted with 1 N HCL), vortexed for 10 min and stirred overnight at 4 °C, protected from light. Pretomanid (Chemshuttle 140130), made fresh every week, was dissolved in 10% 2-hydroxypropyl-β-cyclodextrin (Aldrich 332593) 2% soy lecithin. Vortexed for 10 min and stirred overnight at 4 °C. Moxifloxacin (LTK Laboratories M5794), made fresh every day, was dissolved in water by vortexing for 1 min. Stirred at room temperature until complete dissolution. Clofazimine (LTK Laboratories C458567), made fresh each week, was weighed and transferred to a mortar. The powder was ground with a small volume of 20% 2-hydroxypropyl-β-cyclodextrin (Aldrich 332593), then mixed with 20% 2-hydroxypropyl-β-cyclodextrin. Compound was kept at 4°C and protected from light. Linezolid (LTK Laboratories L3453), made freshly each week, was weighed and transfer to a mortar. The powder was ground with a small amount of PEG-200 (5% of the total liquid volume; Sigma P3015). Transferred to a conical tube and vortexed. 0.5% methylcellulose (95% of the total liquid volume; Sigma M0430) was added and placed in a tube with slow rotation overnight. Compound was kept at 4 °C and protected from light. Pyrazinamide (ACROS 157641000), made fresh each week, was dissolved in sterile ddH2O. The solution was heated to 60 °C until crystals dissolved. Compound was then kept at 4 °C and protected from light. Rifampicin (Sigma R3501), made fresh each week, was ground in a small amount of sterile water. Water was added in small increments while grinding until final volume/concentration was achieved. Compound was kept at 4 °C and protected from light. Ethambutol (Sigma E4630-25G), made fresh weekly, was dissolved in water then kept at 4 °C. Isoniazid (Sigma I3377-50G), was made fresh each week by dissolving in water then kept at 4°C. JNJ-2901 and JNJ-4052 (provided by Janssen Pharmaceutica), were made fresh each week, by dissolving in PEG400 (Aldrich P3265). Compounds were stirred until full dissolution and kept at room temperature. Telacebec (provided by Janssen Pharmaceutica) was made fresh each week by dissolving in PEG400 (Aldrich P3265). 12 N HCl was added (0.36 µL 12 N HCl per mL solution of 2 mg mL−1 telacebec). Compound was stirred overnight and kept at room temperature. In studies A, B, C and D, drugs were administered individually with at least 2 h between dosages. In Study E, individual drugs were made as 2-3x concentrated stocks and combined and dosed together in a 0.2 mL volume at the time of dosing. Specifically: clofazimine and pyrazinamide were made separately at 2 times final concentration at half volume and combined at the time of dosing. Bedaquiline and clofazimine made separately at 2 times final concentration at half volume and combined at the time of dosing. Bedaquiline, clofazimine and pyrazinamide were made separately at 3 times final concentration at one-third volume and combined at the time of dosing. Isoniazid, ethambutol and pyrazinamide were made separately at 3 times final concentration at one-third volume and combined at the time of dosing.

Pharmacokinetics (PK) and tolerability in mouse

The PK of JNJ-2901 and JNJ-4052 was investigated in female Balb-c mice dosed as solution or a suspension at 5 mg kg−1 PO. Three animals were used. Animals had free access to food and water through each study. Blood samples were taken at multiple timepoints up to 24 h after PO dosing. Plasma samples were prepared by protein precipitation with acetonitrile, and the supernatant was analysed for concentrations of compound using a qualified LC-MS/MS method. Individual plasma concentration-time profiles were subjected to a non-compartmental pharmacokinetic analysis (NCA) using Phoenix.

The PK of JNJ-2901 was investigated in female CD-1 mice dosed as solution or a suspension at 5 mg kg−1 PO and JNJ-4052 investigated in male CD-1 mice dosed as solution at 5 mg kg−1 PO. Three anim