Sterilizing Effects of Novel Regimens Containing TB47, Clofazimine, and Linezolid in a Murine Model of Tuberculosis

ABSTRACT

TB47, a new drug candidate targeting QcrB in the electron transport chain, has shown a unique synergistic activity with clofazimine and forms a highly sterilizing combination. Here, we investigated the sterilizing effects of several all-oral regimens containing TB47 plus clofazimine and linezolid as a block and the roles of fluoroquinolones and pyrazinamide in them. All these regimens cured tuberculosis within 4 to 6 months in a well-established mouse model, and adding pyrazinamide showed a significant difference in bactericidal effects.

INTRODUCTION

Tuberculosis (TB), a leading cause of death, resulted in an estimated 10 million new cases and ∼1.4 million deaths in 2019 (1). For the nearly 500,000 annual cases of TB resistant to isoniazid and rifampin (multidrug-resistant TB [MDR-TB]), there are limited treatment options, which have a low treatment success rate (1). Such regimens usually last for about 9 to 12 months and contain at least 4 to 6 drugs, including at least one injectable agent (2). More seriously, TB resistant to isoniazid and rifampin plus other drug(s) needs to be treated longer and results in a higher death rate (3). Considering side effects of injectable drugs and accompanying extensive resistance to fluoroquinolones (FQs) (4–6) and pyrazinamide (Z) (7–10) by drug-resistant TB (DR-TB), we sought to identify a fully oral pan-TB regimen by evaluating the effect of replacing FQ and/or Z with the clinically effective off-patent linezolid (N) (11) in regimens containing a new antimycobacterial candidate, TB47 (T) (12–16), and clofazimine (C), the effect of which further stresses the benefit of dual or multiple targeting of the electron transport chain in enhanced killing of mycobacteria (15, 17–20).

Animal procedures were approved by the Institutional Animal Care and Use Committee of the Guangzhou Institutes of Biomedicine and Health (number 2018053). A logarithmic-phase (optical density at 600 nm [OD₆₀₀], 0.8 to 1.0) broth culture of Mycobacterium tuberculosis H37Rv was used to infect 362 5- to 6-week-old female BALB/c mice by an inhalation exposure system (Glas-Col, Terre Haute, IN, USA). Three mice each from three infection runs were euthanized at day −15 (a day after infection) and day 0 to determine the lung CFU counts. All mice were block randomized prior to treatment initiation. Amikacin (A) was administered subcutaneously, and the other drugs were given by oral gavage (5 days/week), with levofloxacin (L) plus ethambutol (E) or NCT given in combination, at the following doses (in milligrams per kilogram): A, 100; L, 300; E, 100; Z, 150; C, 25; N, 100; and T, 50. Twice the previous dosage of TB47 (16) was used here, as we found that 50 mg/kg showed better activity than 25 mg/kg when used to treat Mycobacterium ulcerans infection (13, 21). Treatment with N was limited to 3 months due to side effects (16), and lung homogenates from mice treated with C-containing regimens were plated on 7H11 plates containing 4% (wt/vol) activated charcoal to overcome carryover effect (22). Bactericidal efficacy was assessed based on lung CFU counts, whereas sterilizing effects were determined by relapse proportions (23). Log-transformed [log₁₀(x + 1)] data were used to determine bactericidal and sterilizing efficacy by two-way analysis of variance and Fisher’s exact test, respectively, using GraphPad Prism version 8.0.2, with P values of <0.05 considered statistically significant.

All nine untreated mice (Table 1) succumbed to TB infection 22 to 25 days postinfection. The log₁₀ CFU/lung were 4.64 ± 0.19 and 8.18 ± 0.14 on day −15 and day 0, respectively. The burden declined significantly in all treated groups during treatment (Fig. 1). All groups remained culture positive after 2.5 months of treatment, except the NCTZ and LNCTZ groups, which had 2/5 and 1/5 culture-negative mice, respectively. L contributed to the bactericidal effect of AEZCT (P = 0.0003) (Fig. 1a) but not to the bactericidal effect of NCT and NCTZ (P > 0.05) (Fig. 1b). Z-containing regimens were more bactericidal than their Z-lacking counterparts (NCT and LNCT, P < 0.0001 and P < 0.001, respectively) (Fig. 1b). ALEZCT was more bactericidal than NCT (P = 0.02), less bactericidal than NCTZ (P = 0.04), and not different from LNCT and LNCTZ (P = 0.097 and 0.65, respectively).

TABLE 1.

Scheme of the experiment

Regimena	No. of mice used to determine lung CFU counts at time pointb
	D−15	D0	M1	M1.5	M2	M2.5	M4	M5	M6	M7	Total
Untreated	9	9	9								27
2ALEZCT/3LEZCT				5	5	5	(15)	(15)			45
2AEZCT/4EZCT				5	5	5	(15)	(15)	(15)		60
3NCT/4CT			5	5	5	5		(15)	(15)	(15)	65
3NCTZ/2CTZ			5	5	5	5	(15)	(15)			50
3LNCT/4LCT			5	5	5	5		(15)	(15)	(15)	65
3LNCTZ/2LCTZ			5	5	5	5	(15)	(15)			50
Total	9	9	29	30	30	30	60	90	45	30	362

^aNumbers in the drug regimen descriptions indicate the number of months (4 weeks/month) for which the drug combination was administered. A, amikacin; L, levofloxacin; E, ethambutol; Z, pyrazinamide; C, clofazimine; N, linezolid; T, TB47. All the drugs were administered once a day, 5 days per week, at the following doses (mg/kg): A, 100; L, 300; E, 100; Z, 150; C, 25; N, 100; and T, 50. A was administered subcutaneously, and the others were administered orally.

^bD−15, 1 day after infection with M. tuberculosis H37Rv; D0, day of treatment initiation; M1, 1 month after treatment initiation, and so on. Numbers in parentheses indicate the number of mice that were held for 4 months after treatment completion before being killed for relapse assessment.

FIG 1.

Bactericidal activities of the studied regimens as observed during the course of experiment. (a) Adding L to AEZCT had a noticeable effect on its bactericidal activity. (b) Addition of L did not result in a significant contribution to activities of any of the NCT-based regimens. However, addition of Z resulted in significant differences between Z-containing and Z-lacking regimens. Incorporation of both L and Z enhanced the activity of NCT, but the activity of NCT with Z alone equals or even surpasses the activity of NCT with both L and Z added. ns, not significant (P > 0.05); ***, P < 0.001; ****, P < 0.0001. D, day; M, month.

Following various durations of different treatments, mice were held for 4 months, which included administration of two doses of 10 mg/kg dexamethasone (24–27) at the beginning of 15th and 16th weeks for immunosuppression prior to relapse assessment (Table 1). Similar to our previous study (16), ALEZCT yielded 13.33% (2/15) and 0% (0/15) relapse in 4 and 5 months, respectively (Table 2), whereas AEZCT (without L) yielded 33.33% (5/15) and 0% (0/15) relapse in 5 and 6 months, respectively (Table 2). Doubling the dose of TB47 did not show enhanced efficacy compared to our previous study (16). So, adding L may shorten the duration from 6 months to 5 months, and the relapse rates at month 5 were significantly different (P = 0.0421). However, an insignificant contribution of L (P > 0.05) was observed between LNCT and NCT at month 5, though LNCT achieved a 0% relapse rate a month earlier than NCT. This was further demonstrated by the consistency in bactericidal and sterilizing effects of NCTZ and LNCTZ. While a shorter course of treatment further corroborates the superior bactericidal effect of NCTZ over NCT (P < 0.0001), the sterilizing effects suggest no significant difference (P > 0.05) (Table 2). We speculate that the course of treatment with NCTZ and LNCTZ could be ≤4 months, but because no mice were examined for relapse in LNCT group at month 4, evidence of a benefit of Z in LNCTZ could not be provided. Nevertheless, all four oral regimens achieved sterilization earlier (4 to 6 months) than or at the same treatment duration as A-containing regimens (5 to 6 months), with NCT achieving sterilization a month longer than ALEZCT. The relapse data (Table 2) suggest that N may be able to replace ALE in the control regimen and still provide potent activity.

TABLE 2.

Proportion of mice relapsing after treatment completion

Regimena	No. of culture-positive mice/total (%) after treatment duration (mo)b
	4	5	6	7
2ALEZCT/3LEZCT	2/15 (13.33) [4, 51]	0/15 (0)	—	—
2AEZCT/4EZCT	5/15 (33.33) [76, 88, 180, 550, 410]	5/15 (33.33) [291, 161, +++, +++, +++]	0/15 (0)	—
3NCTZ/2CTZ	0/15 (0)	0/15 (0)	—	—
3LNCTZ/2LCTZ	0/15 (0)	0/15 (0)	—	—
3NCT/4CT	—	1/15 (6.67) [46]	0/15 (0)	0/15 (0)
3LNCT/4LCT	—	0/15 (0)	0/15 (0)	0/15 (0)

^aNumbers in the treatment regimen descriptions represent the number of months for which the drug combination was administered. Relapse was defined as appearance of one or more M. tuberculosis colonies on the plates spread with the entire lung homogenate 4 months after the indicated duration of treatment. Each lung homogenate was equally divided and plated on 4 plates.

^bThe proportion of mice with culture-positive relapse was determined by holding cohorts of mice for an additional 4 months after treatment completion. Values in brackets indicate the numbers of colonies from lung homogenates of mice with relapse after treatment completion. —, no cohorts of mice were held for relapse assessment after these durations of treatment with these regimens. +++, too many to count.

Limited treatment options for DR-TB create an urgent need for a plethora of safer, accessible, convenient all-oral regimens (28). The WHO even suggests the adoption of a shorter all-oral bedaquiline-containing regimen for treatment of MDR/rifampin-resistant TB (29). Companion agents may determine activities of some anti-TB agents (30), which warrants an emphasis on studying the synergistic activities of novel drugs with no cross-resistance with other anti-TB drugs. Such steps have been taken with the recent FDA’s approval of the combination of bedaquiline, pretomanid, and linezolid (BPaL) (11, 31).

TB47, a novel antimycobacterial candidate (12, 14) that entered good laboratory practice and safety assessment in 2021, has demonstrated good synergy with other anti-TB agents (14–16, 21). A regimen from our previous study (16) was used as a positive control, which is a second-line regimen containing C (22) and T. The use of this second-line regimen is limited by prevalence of resistance to FQs and Z in many DR-TB patients (4, 7), especially in countries with a high TB burden (5, 6). Considering all these factors (4–11, 31), we sought to elucidate the contributions of FQs and/or Z to the regimens, using NCT as a backbone. The NCT-based regimens had good bactericidal and sterilizing effects (Fig. 1b; Table 2), which may cure TB in ≤6 months and even constitute possible all-oral pan-TB regimens. Adding Z to NCT resulted in a marked decline in treatment duration to ≤4 months. However, relapse proportions after 5 months of NCT (1/15) and 4 months of NCTZ (0/15) suggest no significant difference (Table 2). As for L, its bactericidal and sterilizing benefits were limited to ALEZCT (P = 0.0003) (Fig. 1a; Table 2), with only a limited or even no additive effect when added to NCT-containing regimens (P > 0.05), supporting the idea that L and other FQs may be able to be removed from such regimens. However, administration of AEZCT would be subject to extension by 1 month compared to ALEZCT.

In conclusion, we report potentially effective all-oral regimens based on NCT, which could achieve complete sterilization in ≤6 months in a well-established TB mouse model. Though it is too early to infer that their potentials will necessarily translate directly to human TB, the promising results warrant further studies to ascertain their clinical applicability in treatment of DR-TB.