Impact of Dose, Duration, and Immune Status on Efficacy of Ultrashort Telacebec Regimens in Mouse Models of Buruli Ulcer

ABSTRACT

Telacebec (Q203) is a new antituberculosis drug in clinical development that has extremely potent activity against Mycobacterium ulcerans, the causative agent of Buruli ulcer (BU). The potency of Q203 has prompted investigation of its potential role in ultrashort, even single-dose, treatment regimens for BU in mouse models. However, the relationships of Q203 dose, dose schedule, duration, and host immune status to treatment outcomes remain unclear, as does the risk of emergence of drug resistance with Q203 monotherapy. Here, we used mouse footpad infection models in immunocompetent BALB/c and immunocompromised SCID-beige mice to compare different Q203 doses, different dosing schedules, and treatment durations ranging from 1 day to 2 weeks, on long-term outcomes. We also tested whether combining Q203 with a second drug can increase efficacy. Overall, efficacy depended on total dose more than on duration. Total doses of 5 to 20 mg/kg rendered nearly all BALB/c mice culture negative by 13 to 14 weeks posttreatment, without selection of Q203-resistant bacteria. Addition of a second drug did not significantly increase efficacy. Although less potent in SCID-beige mice, Q203 still rendered the majority of footpads culture negative at total doses of 10 to 20 mg/kg. Q203 resistance was identified in relapse isolates from some SCID-beige mice receiving monotherapy but not in isolates from those receiving Q203 combined with bedaquiline or clofazimine. Overall, these results support the potential of Q203 monotherapy for single-dose or other ultrashort therapy for BU, although highly immunocompromised hosts may require higher doses or durations and/or combination therapy.

INTRODUCTION

Treatment of Buruli ulcer (BU), a disease primarily endemic to regions of sub-Saharan Africa and parts of Australia (1), has evolved from extensive surgical excision to the first effective combination chemotherapy regimen of rifampin (RIF) and streptomycin given daily for 8 weeks to the currently recommended oral regimen of RIF and clarithromycin (CLR) for 8 weeks (2, 3). Although clinical studies have shown good efficacy of the RIF-CLR regimen (4), shortening the duration of treatment and reducing the potential for adverse effects and drug-drug interactions would make it easier to implement (5–8).

Telacebec (Q203), a new drug in clinical development to treat tuberculosis (9), is extremely potent against Mycobacterium ulcerans, the causative agent of BU, due to the high vulnerability of the bacterial respiratory cytochrome bc₁:aa₃ oxidase to this agent (9) in the absence of a functional alternate terminal oxidase (10). In our previous experiments, we demonstrated the efficacy of Q203 in a mouse footpad infection model of BU, alone and in combination with rifapentine and/or other drugs acting on the respiratory chain, including its potential to drastically reduce the treatment duration needed for cure (11, 12). Q203 at doses as low as 0.5 mg/kg given alone for 5 days showed remarkable bactericidal activity, extending well beyond the last day of drug ingestion (11). This, coupled with reports from other investigators, has led to suggestions to explore the possibility of a simple one-dose treatment of BU (11, 13, 14). However, several important questions remain to be answered before optimized Q203 regimens can be recommended for clinical study. For example, is the same total dose of Q203 as effective when given as a single dose compared to daily divided doses, and what is the optimal duration? Second, would Q203 monotherapy lead to selection of drug resistance? Finally, can addition of a second drug increase the efficacy of monotherapy and/or prevent emergence of Q203 resistance?

We designed the current experiments in the mouse footpad infection model to answer these questions. First, we performed a dose-ranging and dose-fractionating study of Q203 to determine the total dose necessary for cure and whether a single dose or divided daily doses is a more effective way to deliver the same total dose. Second, to evaluate the ability of Q203 monotherapy to sterilize infected footpads without selecting drug-resistant mutants in mice with deficient innate and adaptive immune responses, we included SCID-beige mice lacking adaptive T-cell and B-cell responses and functional natural killer cells in selected treatment groups. In our previous study, mice were observed for a maximum of 4 weeks posttreatment. Here, we extended the observation period up to 16 weeks posttreatment to give a longer time for mice to relapse if viable bacteria remained in the footpads after treatment. Third, we also investigated if two-drug regimens combining Q203 with bedaquiline (BDQ), clofazimine (CFZ), or CLR increase efficacy over that of Q203 alone and/or prevent the selection of Q203-resistant mutants in both BALB/c and SCID-beige mice.

Our results demonstrate that single-dose treatment may indeed be as effective as daily divided doses and successfully resolve BU. However, Q203 was less effective and selected for drug-resistant mutants in SCID-beige mice, indicating that monotherapy may not be appropriate for use in highly immunocompromised hosts or that such hosts may benefit from higher doses, longer durations of treatment, or intensified combination therapy.

RESULTS

Experiment 1: evaluating the impact of dose, schedule, duration, and immune status on the efficacy of Q203.

In this experiment, BALB/c mice and SCID-beige mice with established footpad infections were treated with either the RIF-CLR (RC) control regimen or one of the test regimens shown in Table S1 in the supplemental material. Four total doses of Q203 were tested in BALB/c mice, as follows: 2.5, 5, 10, and 20 mg/kg. The 2.5 mg/kg total dose was given as a single dose or was divided into five consecutive daily doses of 0.5 mg/kg. The 5 mg/kg total dose was given as 10 daily (5 days per week [5/7]) doses of 0.5 mg/kg over 2 weeks. The 10 mg/kg total dose was given as a single dose or was divided into three consecutive daily doses of 3.3 mg/kg or five daily doses of 2 mg/kg. The 20 mg/kg total dose was given as 10 daily (5/7) doses of 2 mg/kg over 2 weeks. SCID-beige mice received either RIF-CLR, 0.5 mg/kg Q203 (Q_0.5), or 2 mg/kg Q203 (Q₂) for one or 2 weeks (total doses of 2.5, 5, 10, and 20 mg/kg). Single-dose regimens were not included for testing in SCID-beige mice.

(i) Footpad swelling and CFU counts in BALB/c mice.

The mean (± standard deviation [SD]) CFU count on the day after footpad infection was 3.97 ± 0.52 log₁₀ per footpad. Mice exhibiting footpad swelling grades between 2 and 3 at 8 weeks postinfection were included in the study and were randomized to treatment arms. At the start of treatment (day zero [D0]), the median swelling grade was 2.5. Q203 rapidly reduced swelling in all dosing groups to a median swelling grade of ≤1 within the first week of treatment and to a negligible swelling grade by weeks 3 to 5, irrespective of whether treatment was ≤1 week (Fig. 1A) or 2 weeks (Fig. 1B) in duration. Mice receiving RIF-CLR experienced a slower but persistent improvement in swelling after completing 1 or 2 weeks of treatment, and those completing 2 weeks reached a state of negligible swelling between weeks 6 and 8, but the median increased back to 0.5 by week 14, as it did for the group that received a single 2.5 mg/kg dose of Q203.

FIG 1.

Footpad swelling grade and microbiological outcome in BALB/c mice in response to treatment. Median swelling grade in BALB/c mice treated for 1 week (A) and BALB/c mice treated for 2 weeks (B). Log₁₀ CFU/footpad in BALB/c mice treated for 1 week (C) and BALB/c mice treated for 2 weeks (D). Solid lines represent change during treatment, while that after stopping treatment is shown by dashed lines. RC, rifampin (RIF) 10 mg/kg plus clarithromycin (CLR) 100 mg/kg. Numbers in subscripts after the drug abbreviation are the doses in mg/kg. D0, day 0 or the beginning of treatment. The lower limit of detection (LLOD) at week 1 was 1.48 log₁₀ CFU. At week 2, it was 0.48 log₁₀ CFU. For other time points, it was 0.40 log₁₀ CFU.

The mean CFU count at D0 was 6.12 ± 0.55 log₁₀ per footpad (Fig. 1C and D). Mean CFU counts in RIF–CLR-treated mice declined in a duration-dependent manner during treatment and over the next 4 weeks posttreatment before reaching a plateau or increasing again by 13 weeks posttreatment. Two out of 6 footpads were culture negative 4 weeks after completing 2 weeks of RIF–CLR treatment, but all 8 footpads were positive at 13 weeks posttreatment, with a mean CFU count of nearly 5 log₁₀ (Table S2). As suggested by the swelling results, Q203 was more effective. Just 1 week (i.e., 5 doses) of Q_0.5 rendered 2 out of 6 and 2 out of 8 footpads culture negative at 4 and 13 weeks posttreatment, respectively. Extending this dose for another week rendered all 8 footpads culture negative at 13 weeks later, except for 1 footpad with 1 CFU. Overall, 5 doses of Q_0.5 were significantly more effective than 10 doses of RIF-CLR when assessed at week 1 + 4 (P < 0.0001), but not at week 1 + 13. Ten doses of Q_0.5 were significantly more effective than 10 doses of RIF-CLR when assessed at weeks 2 + 4 and 2 + 13 (P = 0.015 and P < 0.001, respectively). The higher daily dose of Q₂ for 1 or 2 weeks rendered all footpads culture negative at 4 and 13 weeks posttreatment, except for a single footpad with 1 CFU at week 1 + 13. While superior to RIF-CLR (P < 0.01 at weeks 1 + 4, 1 + 13, 2 + 4, and 2 + 13), Q₂ was not statistically superior to Q_0.5 at these time points. A single dose of 2.5 mg/kg Q203 (Q_2.5) resulted in a higher mean CFU/footpad count at week 1 + 13 than the same total dose divided over 5 days, but the difference was not statistically significant. A single dose of 10 mg/kg Q203 (Q₁₀) or 3 daily doses of 3.3 mg/kg Q203 (Q_3.3) appeared to be just as effective as the same total dose divided over 5 days (Q₂). Each of these regimens resulting in a total dose of 10 mg/kg was superior to RIF-CLR at week 1 + 4 (P < 0.0001) and week 1 + 13 (P < 0.05) and to the single 2.5 mg/kg dose at week 1 + 13 (P < 0.05).

(ii) Footpad swelling and CFU counts in SCID-beige mice.

The mean CFU count the day after footpad infection was 4.17 ± 0.22 log₁₀ CFU/footpad. At D0, the median swelling grade was 2.6 (Fig. 2A and B), and mice were randomized to treatment arms. As observed in BALB/c mice, both Q203 doses rapidly reduced the swelling grade, which became negligible within 1 to 2 weeks of completing 1 to 2 weeks of treatment with either dose and remained negligible to week 14 except in the group receiving Q_0.5 for 1 week; mice in this group uniformly experienced recrudescence of swelling at week 10 and required euthanasia thereafter. RIF-CLR reduced the swelling grade less rapidly than Q203 and only while mice were on treatment, as footpad swelling rapidly rebounded within 1 to 3 weeks of stopping RIF-CLR treatment and eventually necessitated euthanasia.

FIG 2.

Footpad swelling grade and microbiological outcome in SCID-beige mice in response to treatment. Median swelling grade in SCID-beige mice treated for 1 week (A) and SCID-beige mice treated for 2 weeks (B). Log₁₀ CFU/footpad in SCID-beige mice treated for 1 week (C) and SCID-beige mice treated for 2 weeks (D). Solid lines represent change during treatment, while that after stopping treatment is shown by dashed lines. RC, RIF 10 mg/kg plus CLR. Numbers in subscript after the drug abbreviation are the doses in mg/kg. D0, day 0 or the beginning of treatment. Lower limit of detection (LLOD) at week 1 was 1.48 log₁₀ CFU, at week 2 was 0.48 log₁₀ CFU, and for other time points, it was 0.40 log₁₀ CFU.

The mean CFU count at D0 was 6.56 ± 0.29 log₁₀ per footpad (Fig. 2C and D). RIF-CLR treatment resulted in duration-dependent reduction in footpad CFU counts of up to 1.3 log₁₀ after 2 weeks, but the CFU counts increased again to approximately 6 log₁₀ by the time the humane endpoint was reached at week 2 + 5. Q_0.5 for 1 week resulted in a greater bactericidal effect (i.e., 3.0 log₁₀ reduction) at week 1 + 4, but the CFU counts again approached 6 log₁₀ when euthanasia was required at week 1 + 9. Increasing the total dose by extending the duration of Q_0.5 to 2 weeks or giving Q₂ for 1 or 2 weeks prevented such treatment failures and reduced the CFU counts at 4 weeks posttreatment in a dose- and duration-dependent fashion. Unlike in BALB/c mice, Q₂ was superior to Q_0.5 at both week 1 + 4 and week 2 + 4 (P < 0.05 and P < 0.0001, respectively) in SCID-beige mice. At 13 weeks posttreatment, between 4 and 5 of the 8 footpads in each group were culture negative, suggesting that the pathogen had been cleared from at least half of the mice.

Despite its powerful bactericidal and sterilizing activity in both mouse strains, Q203 was less potent in SCID-beige mice. BALB/c mice had significantly lower mean CFU counts than those of SCID-beige mice after treatment with Q_0.5 (P < 0.05 at weeks 1 and 2; P < 0.001 at weeks 1 + 4 and 2 + 4) and Q₂ (P = 0.001 at week 1 + 4).

(iii) Resistance detection.

By plating serial 10-fold dilutions of a culture suspension of M. ulcerans on agar plates supplemented with Q203 at 4 and 40 times the previously determined MIC of 0.075 ng/ml (12), the frequency of spontaneous Q203-resistant mutants in vitro was determined to be 1.2 in 10⁵ on plates containing Q203 at 0.3 ng/ml (i.e., 11 CFU per 900,000 total CFU plated). No CFU were detected on plates containing Q203 at 3 ng/ml, indicating that the frequency of resistance at 40× MIC was less than 1 in 10⁶. In a second experiment with a higher bacterial inoculum (6 × 10⁷ CFU/ml), the frequency of resistance to 3 ng/ml was 5 in 10⁸.

Isolates from 4 BALB/c mouse footpads obtained at 13 weeks posttreatment from mice that received 5 doses of Q₂ (n = 1), a single dose of Q₁₀ (n = 2), or 10 doses of Q_0.5 (n = 1) were plated in serial dilutions on agar containing 0.3 and 3 ng/ml of Q203, and no Q203-resistant mutants were detected. Isolates from 2 SCID-beige mice among those receiving 5 doses of Q_0.5 that uniformly experienced early footpad reswelling had approximately 1 in 10⁵ CFU growing at 3 ng/ml and 1 in 10⁴ CFU growing at 0.3 ng/ml, indicating that, although some selective amplification of resistant mutants may have occurred with treatment, this was not responsible for the treatment failure. In SCID-beige mice treated for 2 weeks with Q_0.5, 4 of 8 footpads remained culture positive, and no selective amplification of resistant mutants was detected on either 0.3 or 3 ng/ml. However, treatment with Q₂ for 2 weeks did select Q203-resistant mutants in some SCID-beige mice. At 13 weeks postinfection, 3 of 8 footpads were culture positive, and all 3 isolates were resistant to Q203 at both 0.3 and 3 ng/ml. Three colonies from each of these 3 isolates were subjected to PCR amplification and sequencing of the qcrB gene. All isolates had the same mutation, a single-nucleotide polymorphism (SNP) (A → G) at nucleotide (nt) 967 resulting in the amino acid change T323A.

Experiment 2: determining if addition of a companion drug increases the sterilizing activity of Q203 and finding the shortest possible treatment duration.

In this experiment, BALB/c mice were treated with Q₂ alone or in combination with one of the following drugs: 25 mg/kg BDQ (Q₂BDQ₂₅), 12.5 mg/kg CFZ (Q₂CFZ_12.5), or 100 mg/kg CLR (Q₂CLR₁₀₀). Drugs were administered once daily via oral gavage, and each regimen was given for 1, 3, and 5 days. Cohorts of mice were followed for up to 15 weeks after the start of treatment. The scheme of the experiment is shown in Table S3.

(i) Footpad swelling and CFU counts up to 5 weeks posttreatment.

On the day after infection, the mean (±SD) CFU count was 4.95 ± 0.38 log₁₀ CFU/footpad. At the start of treatment (D0), 34 days after infection, the mean CFU count was 6.62 ± 0.34 log₁₀ CFU/footpad, and the median swelling grade was 2 on a scale of 0 to 4 (15). Untreated mice subsequently exhibited increased footpad swelling and required euthanasia during the experiment (Fig. 3A), with the exception of one untreated mouse that survived until the end of the experiment without reaching the endpoint for euthanasia despite harboring 6.17 ± 0.86 log₁₀ CFU/footpad at the end of the experiment. In mice treated with R₁₀CLR₁₀₀ for 10 consecutive days, the swelling grade decreased slightly to an average of 1.7 after 1 week of treatment and to 1.0 after 2 weeks. The decrease in swelling continued up to week 4 and was maintained until week 8, when a gradual increase in swelling started and ultimately resulted in four footpads having a swelling grade of ≥3 by week 14. The mean CFU count in R₁₀CLR₁₀₀-treated mice decreased to 3.81 ± 1.35 log₁₀ CFU/footpad after 10 days of treatment (Fig. 3B) but increased after the end of treatment to 5.49 ± 1.21 log₁₀ CFU/footpad by week 15.

FIG 3.

Footpad swelling grade and microbiological outcome in BALB/c mice in response to treatment. (A) Median swelling grade in BALB/c mice during and after treatment, with the indicated regimens and number of doses in parentheses. (B) Log₁₀ CFU/footpad 1 week after the start of treatment (or 12 days for R₁₀CLR₁₀₀). (C) Log₁₀ CFU/footpad 6 weeks after the start of treatment. (D) Log₁₀ CFU/footpad 15 weeks after the start of treatment. R, rifampin, CLR, clarithromycin, Q, telacebec, BDQ, bedaquiline, CFZ, clofazimine. Numbers in subscripts indicate the dose in mg/kg. D0, day 0 or the beginning of treatment. Horizontal bars indicate medians.

As previously observed (16), Q203-containing regimens rapidly decreased footpad swelling, resulting in >98% of these mice having a swelling grade of ≤1 by week 1, significantly lower than that in the R₁₀CLR₁₀₀ control group (P < 0.001). Among mice receiving Q203, 36% and 67% of the footpads showed no swelling by weeks 2 and 4, respectively, with no significant difference between Q203-treated groups. In the first week, the superior effect of Q203 was not yet reflected in footpad CFU counts. Five weeks after the end of treatment, all Q203-containing regimens given for 5 days resulted in culture-negative footpads (Fig. 3C). With the exception of 3 of 6 footpads from mice treated for 3 days with Q₂CFZ_12.5, footpads treated with a Q203-containing regimen for 3 days were all culture negative. In the groups treated for only 1 day, 1 mouse (both footpads) that received Q203 alone had positive culture results, all mice receiving Q₂BDQ₂₅ for 1 day were culture negative, one mouse receiving Q₂CFZ_12.5 for 1 day had a culture-positive footpad, and in the group receiving Q₂CLR₁₀₀ for 1 day, two mice had both footpads culture positive at this time point. At week 6, the CFU counts were significantly higher in mice treated with one dose of Q₂CLR₁₀₀ compared those in mice treated with one dose of Q₂BDQ₂₅ (P < 0.03).

Eight BALB/c mice from each treatment group were held for 15 weeks to evaluate relapse. Apart from the relapses observed in all R₁₀CLR₁₀₀-treated mice, we also observed relapse in all but 3 of the 16 footpads treated with a single dose of Q₂CLR₁₀₀. A single dose of Q203 resulted in 9 out of 16 footpads being culture negative, a single dose of Q203 and BDQ resulted in half (8/16) of the footpads being culture negative, and one dose of Q₂CFZ_12.5 resulted in 7 of 16 footpads being culture negative. At week 15, the CFU counts were significantly higher in mice treated with one dose of Q₂CLR₁₀₀ compared to one dose of Q203 alone, Q₂BDQ₂₅, or Q₂CFZ_12.5 (P < 0.02). When given for 3 days, Q203 with or without a companion drug prevented relapse in all footpads with the exception of 2 out of 16 footpads from mice treated with Q₂CLR₁₀₀. Q₂ alone or with a companion drug given for 5 days prevented relapse in all footpads (Fig. 3D). The limit of detection was 1 CFU.

Experiment 3: determining the impact of the host immune response on the sterilizing activity of Q203-containing regimens and the selection of Q203-resistant mutants.

This experiment was conducted to determine if a 5-day treatment with Q₂ daily, alone or in combination with 25 mg/kg BDQ (Q₂BDQ₂₅) daily or 12.5 mg/kg CFZ (Q₂CFZ_12.5) daily can successfully treat M. ulcerans footpad infection in SCID-beige mice without selection of Q203-resistant mutants. The scheme of the experiment is shown in Table S4.

(i) Footpad swelling and CFU count results.

One day after infection, the mean (±SD) CFU count was 4.17 ± 0.22 log₁₀ CFU/footpad. At the start of treatment (D0), 40 days after infection, the median footpad swelling grade was 2.5 (Fig. 4A), and the mean CFU count was 6.26 ± 0.35 log₁₀ CFU/footpad for the mice allocated to the Q203-containing regimens and 6.56 ± 0.29 log₁₀ CFU/footpad for the mice allocated to the R₁₀CLR₁₀₀ control group (Fig. 4B). In all Q203-containing treatment groups, a sharp reduction in footpad swelling to a median swelling grade of 0.25 after 1 week of treatment was observed; in the Q₂CFZ_12.5 group, 19 of 34 footpads had a swelling grade of 0, and in the Q₂ and Q₂BDQ₂₅ groups, 7 out of 34 and 8 out of 34 footpads, respectively, had a swelling grade of 0. In contrast, the swelling of untreated mouse footpads continued to increase such that the mice reached the humane endpoint for euthanasia at week 1, while the control R₁₀CLR₁₀₀ regimen decreased footpad swelling after 1 week to a median swelling grade of only 1.5, which was still significantly higher than that in mice given the Q203-containing regimens (P < 0.001). At 1 week after the end of treatment, in every group receiving Q203 the median swelling grade was 0, whereas the swelling started to increase in R₁₀CLR₁₀₀-treated mice again, reaching median swelling grades of 2.2 at week 2 and ≥2.5 at week 3, mandating that all R₁₀CLR₁₀₀-treated mice be euthanized per protocol at week 4. Footpads in the Q203-treated groups maintained a swelling grade of approximately 0 until the end of the experiment. At week 3, only 2 of the 24 footpads from mice treated with a Q203-containing regimen were culture negative.

FIG 4.

Footpad swelling grade and microbiological outcome in SCID-beige mice in response to treatment. (A) Median swelling grade in SCID-beige mice during and after treatment with the indicated regimens and number of doses in parentheses. (B) Log₁₀ CFU/footpad in SCID-beige mice in the indicated regimens during the study. (C) Log₁₀ CFU/footpad 5 weeks after the start of treatment. (D) Log₁₀ CFU/footpad 17 weeks after the start of treatment. R, rifampin, CLR, clarithromycin, Q, telacebec, BDQ, bedaquiline, CFZ, clofazimine. Numbers in subscripts indicate the dose in mg/kg. D0, day 0 or the beginning of treatment. Horizontal bars indicate medians.

Trends in footpad CFU counts largely mirrored the swelling grades, although there was not as much discrimination between the R₁₀CLR₁₀₀ control group and the Q203-treated groups at week 1, and the minimum values were not reached in Q203-treated groups until week 5 (4 weeks after treatment completion) or later. At 3 weeks after the end of treatment, when all R₁₀CLR₁₀₀-treated mice were euthanized per protocol, their footpad CFU counts had still not decreased and were near the pretreatment baseline. In contrast, in Q203-treated mice, footpad CFU counts continued to decline after dosing ended and, by week 5, all footpads in the Q₂BDQ₂₅-treated group were culture negative; in the Q₂CFZ_12.5 group, two footpads (from two different mice) were culture negative, and in the Q203 alone group, three footpads (including both footpads in one mouse) were culture negative (Fig. 4C). The differences between the Q₂BDQ₂₅, the Q203 alone, and the Q₂CFZ_12.5 groups were statistically significant (P < 0.03). The median footpad swelling grade, as well as the mean CFU count in mice treated with any Q203-containing regimen, was significantly (P < 0.0001) lower than those in the R₁₀CLR₁₀₀ control group. The limit of detection was 3 CFU per footpad.

Five SCID-beige mice of each Q203-containing treatment group were held for 17 weeks (16 weeks after treatment completion). No mouse showed a relapse in footpad swelling. In the group receiving Q203 alone, 2 out of 10 footpads had detectable CFU; in the Q₂CFZ_12.5 group, one footpad had detectable CFU, and in the Q₂BDQ₂₅ group, 2 of 10 footpads had 1 CFU each (limit of detection, 1 CFU) (Fig. 4D).

(ii) Resistance detection.

The M. ulcerans isolates from the eight SCID-beige footpads with the highest CFU counts at week 5 were selected for Q203 susceptibility testing. Two of the four footpad isolates from mice receiving Q203 alone showed resistance (40% and 2.2% of total CFU on 3 ng/ml, 40% and 3.9% of total CFU on 0.3 ng/ml, respectively) (Table S5). The Q203-containing plates from two Q₂CFZ_12.5-treated footpads were contaminated and were not assessable, whereas the isolates from the other two footpads from Q₂CFZ_12.5-treated mice did not show resistance.

DISCUSSION

The standard treatment for BU is an 8-week course of RIF-CLR (17). A shorter regimen remains a high priority in the search for a better treatment (13, 18). Q203 is a new drug targeting bacterial respiration that is in clinical development for tuberculosis and has also proven to be very effective against M. ulcerans in preclinical studies (11–14, 16). Previously, we showed its potential for reducing the duration of treatment to 1 to 2 weeks, based in large part on the extended duration of bactericidal activity after stopping treatment (11). At least a proportion of this extended activity of Q203 is attributable to its potent activity and slow clearance from plasma and the site of infection. However, we also questioned whether treatment with Q203 might promote bacterial clearance by host-mediated immune mechanisms. An effective host response may also have an important role in preventing the selection of drug-resistant bacteria, especially if Q203 monotherapy were being considered. In the current study, we explored the possibility of a single-dose or other ultrashort treatment duration for M. ulcerans compared to 1 to 2 weeks of treatment and assessed the impact of dose, dosing schedule, and the host immune response on the efficacy of such regimens by comparing long-term posttreatment outcomes in immunocompetent BALB/c mice and immunocompromised SCID-beige mice. This also afforded the opportunity to assess the potential for emergence of drug resistance during monotherapy with Q203 and the impact of host immunity.

Q203 again exhibited strong bactericidal effects that extended well beyond the end of the dosing period, even if that dosing period consisted of a single dose. In BALB/c mice, footpad swelling resolved almost completely in as little as 1 to 2 weeks after treatment, including in groups treated with single-dose regimens. Q203 showed clear dose-ranging activity, with higher doses showing increased activity. Total doses of 5 to 20 mg/kg achieved relapse-free cure in the vast majority of footpads at 13 weeks posttreatment, while total doses of 2.5 mg/kg were less effective. A single dose of 10 mg/kg was equally effective as 2 mg/kg given daily for 5 days or 3.3 mg/kg for 3 days. While our study was under way, Thomas et al. (13) reported that a single Q203 dose of 20 mg/kg given once or divided into four doses of 5 mg/kg given over a week resulted in relapse-free cure in a mouse footpad model. Our study supports and extends those results by showing that a lower single-dose regimen of Q203 at 10 mg/kg or a 6 to 10 mg/kg total dose divided into three daily doses of 2 to 3.3 mg/kg were also highly effective. Our prior pharmacokinetics (PK) data (11) suggest that Q203 doses between 2 and 10 mg/kg likely correspond well to human doses of 100 to 300 mg that were recently reported to be well tolerated and safe in phase 1 trials and in tuberculosis patients over 14 days of dosing in a recent phase 2a trial (19). Although a total dose of 20 mg/kg given over a period of 2 weeks given as 2 mg/kg daily, 5 days per week, was the only regimen that eliminated all cultivable bacilli in all BALB/c mouse footpads in experiment 1, 10 mg/kg divided over 5 days left only a single culture-positive footpad with a single detectable CFU in experiment 1, and 6 to 10 mg/kg divided over 3 to 5 days rendered all footpads negative in experiment 2. This is the lowest total dose of Q203 reported to render all footpads culture negative in a mouse M. ulcerans footpad infection model. Taken together, these results provide strong support for the prospect of single-dose or other ultrashort therapy of M. ulcerans infection with doses that have thus far been safe and well tolerated in humans. In January 2021, the U.S. Food and Drug Administration granted an orphan drug designation to Q203 for BU treatment (http://www.koreabiomed.com/news/articleView.html?idxno=10169), providing further incentives for clinical development of Q203 for this indication.

Comparisons of regimen efficacy between BALB/c and SCID-beige mice demonstrated important effects of the host immune response on the response to treatment with both Q203 and the RIF-CLR control regimen. BALB/c mice treated with RIF-CLR for 1 or 2 weeks experienced a gradual reduction in footpad swelling and CFU for at least 4 weeks after stopping treatment and did not have recrudescent swelling of the footpad, despite the vast majority of the footpads remaining culture positive with >1,000 CFU at 13 weeks posttreatment. In stark contrast, SCID-beige mice experienced a rebound in footpad swelling almost immediately after stopping treatment, as well as little additional decline in footpad CFU counts in the first few weeks posttreatment before the burden increased again with longer incubation (i.e., in the group treated for 2 weeks). These results indicate that the magnitude and duration of the bactericidal effect of RIF-CLR is, to a significant extent, dependent on effective natural killer cells and/or adaptive host immunity, as is the posttreatment resolution of footpad swelling and containment of bacterial growth. These immune effects may be enhanced by reductions in mycolactone synthesis resulting from fewer viable bacteria available to produce mycolactone and possibly from nonspecific drug effects on the expression of the enzymatic machinery responsible for mycolactone production.

The efficacy of Q203 also depended on mouse strain. Although resolution of swelling during Q203 treatment was at least as rapid in SCID-beige mice as that in BALB/c mice, the rate of bacterial killing and the extent of bacterial eradication were not as great in SCID-beige mice. Nevertheless, sustained bactericidal effects were observed after Q203 treatment in SCID-beige mice, and the majority of those treated with Q203 at 2 mg/kg for 1 to 2 weeks were culture negative at 13 to 15 weeks posttreatment, despite having mean CFU counts of 4.5 to 5.7 log₁₀ at the end of dosing. This striking result affirms the persistent bactericidal effects and sterilizing efficacy of Q203, even in the absence of an effective adaptive host immune response. However, the differences between mouse strains also confirm the contribution of the immune response in aiding the successful clearance of M. ulcerans. Our results suggest that treatment of BU in immunosuppressed patients, for example, in the case of advanced HIV coinfection or patients undergoing iatrogenic immunosuppression, may need to be adjusted to reach a complete cure in all patients when using Q203.

To our knowledge, this study is the first to evaluate if two-drug combinations of Q203 with BDQ, CFZ, or CLR could shorten the duration of treatment compared to that of Q203 monotherapy. No tested combination including Q203 at 2 mg/kg rendered all footpads culture negative after a single dose. For this purpose, higher doses or other companion drugs may be necessary to achieve a single-dose cure. In BALB/c mice, none of the combinations tested was significantly more effective than Q203 alone. In SCID-beige mice, at week 5 of experiment 3, the difference between Q₂BDQ₂₅, which rendered all footpads culture negative; Q203 alone, which rendered 3 of 8 footpads culture negative; and Q₂CFZ_12.5, which rendered 2 of 8 footpads culture negative, was statistically significant. However, at the end of the follow-up period, the difference in CFU counts was not significant.

Another important conclusion drawn by comparing Q203 treatment responses in BALB/c and SCID-beige mice is that the host immune response plays a key role in preventing the emergence of Q203 resistance during treatment. Monotherapy of active mycobacterial infections is generally strongly discouraged because of the propensity for selecting spontaneous drug-resistant mutants. However, as we and others have argued previously (11, 13, 16), Q203 monotherapy of BU may be considered given the low spontaneous frequency of resistance mutations in M. ulcerans, their potential fitness cost, the potential for adaptive host immunity to contain or clear small residual populations of drug-resistant bacteria, and the small-to-absent risk of person-to-person transmission of a drug-resistant infection even if it occurs. As we hoped, we observed no evidence of Q203 resistance among the isolates obtained from the last few culture-positive BALB/c mice at 13 weeks posttreatment. In contrast, all SCID-beige mice remaining culture positive after 2 weeks of treatment with Q203 at 2 mg/kg in experiment 1 harbored Q203-resistant isolates with a mutation in the qcrB gene encoding the drug target. This resistance mutation was previously reported by Scherr et al. to result in a 230.5-fold increase in the MIC of Q203 (16), which is consistent with growth of our isolates on agar containing Q203 at 40× MIC. The frequency of spontaneous Q203-resistant mutants was previously reported to be 1 in 10⁹ at 10 nM (5.57 ng/ml) (16), while we found it to be around 1 in 10⁵ at 0.3 ng/ml and 5 in 10⁸ at 3 ng/ml in our in vitro selection studies. As the CFU counts at the start of treatment in SCID-beige mice were 6.56 ± 0.29 log₁₀ CFU per footpad, it is possible that some of the mice harbored spontaneous Q203-resistant mutants at the start of treatment and that these were selectively amplified at a higher dose of Q203 at 2 mg/kg under its strong selective pressure in the absence of an effective host immune response ultimately resulting in treatment failure. No resistance was observed when Q203 was combined with CFZ in experiment 3. Although only a limited number of isolates was assessed, this finding suggests that a companion drug might help to prevent emergence of Q203 resistance in immunocompromised hosts. It should be recognized that SCID-beige mice represent an extreme state of immunosuppression, and the majority of these mice receiving monotherapy appeared to be cured. Therefore, the pros and cons of monotherapy versus combination therapy will require careful consideration and clinical evaluation. The overall positive outcomes of monotherapy in SCID-beige mice, and its simplicity, reduced cost, and low potential for adverse effects, as well as the very low likelihood of human-to-human transmission of Q203-resistant BU (20, 21), should it occur, seem to favor monotherapy. Nevertheless, further investigations may be warranted to identify an optimal companion drug that could better ensure the efficacy of ultrashort (i.e., 1- to 3-dose) regimens across the spectrum of host immune competency (15, 22).

In summary, we identified the lowest-yet-published total dose of Q203 leading to a complete healing of all affected footpads in a mouse model of BU, further paving the way toward clinical evaluation of efficacy in M. ulcerans-infected patients. Surprisingly, none of the companion drugs evaluated contributed an additional treatment-shortening effect in BALB/c mice, nor did they lead to a higher overall rate of cure in SCID-beige mice. Our study confirms and extends prior observations of the dramatic treatment-shortening potential of Q203 for BU but also raises the question of the optimal treatment for immunosuppressed individuals.

MATERIALS AND METHODS

Bacterial strain.

M. ulcerans strain 1059, originally obtained from a patient in Ghana, was used for this study (23).

Antibiotics.

RIF and CFZ were purchased from Sigma. CLR pills were purchased from the Johns Hopkins Hospital pharmacy. Q203 and BDQ were kindly provided by the TB Alliance. RIF, CLR, and CFZ were prepared separately in sterile 0.05% (wt/vol) agarose solution in distilled water. Q203 was formulated in 20% (wt/wt) d-α tocopheryl polyethylene glycol 1000 (Sigma) succinate solution (11). BDQ was formulated in acidified 20% hydroxypropyl-β-cyclodextrin solution.

Mouse infection.

To prepare the inoculum, mouse footpads infected with M. ulcerans 1059 were harvested upon reaching a swelling grade between 2 and 3. After thorough disinfection with 70% alcohol swabs, the footpad tissue was dissected away from bone and then homogenized by fine mincing before being suspended in sterile phosphate-buffered saline (PBS). The suspensions were then frozen in 1.5-ml aliquots and stored at −80°C. Prior to infection, vials were thawed to room temperature and pooled to obtain the required amount for mouse infection. BALB/c and Fox Chase SCID Beige mice (Charles River Laboratories) were inoculated subcutaneously in both hind footpads with 0.03 ml of a culture suspension containing M. ulcerans 1059. Treatment began approximately 8 weeks (experiment 1) or 5 to 6 weeks (experiments 2 and 3) after infection, when the mice had a footpad swelling grade of 2 or 3. The CFU counts at implantation were determined from six footpads from three mice on the day after infection and at the start of treatment (D0) to determine the infectious dose and the pretreatment CFU counts, respectively.

Treatment.

Drugs were administered orally in 0.2 ml by gavage. Drug doses were chosen to produce similar area under the plasma concentration-time curves over 24 h postdose compared to those for human doses, as previously described (11, 12, 15). All animal procedures were conducted according to relevant national and international guidelines and approved by the Johns Hopkins University Animal Care and Use Committee. Mice were randomized to treatment groups (see Table S2 in the supplemental material). The control regimen consisted of RIF 10 mg/kg plus CLR 100 mg/kg. Test regimens consisted of either Q_0.5 or Q₂ given 5 days per week (5/7) for 1 to 2 weeks, Q_3.3 given daily for 3 consecutive days, or Q_2.5 or Q₁₀ given as a single dose. The single dose of Q_2.5 was selected to match the total dose of Q_0.5 given for 1 week, and the Q₁₀ single dose matched the total doses of Q₂ given for 1 week and Q_3.3 given for 3 days. Combination regimens included Q₂ plus either CLR 100 mg/kg, BDQ 25 mg/kg, or CFZ 12.5 mg/kg. SCID-beige mice were included to provide information about the sterilizing potential of regimens compared to the activity in BALB/c mice and the risk of selection of drug resistance during monotherapy in an immunocompromised host.

Evaluation of treatment response.

Treatment outcomes were evaluated based on (i) change in footpad swelling, denoted as swelling grade, and (ii) change in footpad CFU counts. The swelling grade was scored as described previously (15). Briefly, the presence and the degree of inflammatory swelling of the infected footpad were assessed weekly and scored from 0 (no swelling) to 4 (inflammatory swelling extending to the entire limb) for all surviving mice. For CFU counts, six footpads (from three mice) were evaluated on the day after infection and at the start of treatment (D0) to determine the infectious dose and the pretreatment CFU counts, respectively. The response to treatment was determined by plating 6 footpads (from 3 mice) at the end of 1 week or 2 weeks of treatment. Mice treated with a single dose or the 3-day regimen were sacrificed at the end of 1 week along with 1-week treated mice. As shown previously (11), Q203 exhibits extended activity after stopping treatment. To evaluate this, mice from 1-week and 2-week treatment groups were held for additional predetermined periods of up to 17 weeks without treatment. During this period, footpads were inspected every 2 weeks for any signs of reswelling after stopping treatment. When reswelling was observed, mice were sacrificed when the swelling reached a lesion index of ≥3, and the footpads were harvested and plated for CFU counts. All of the remaining mice were sacrificed at the end of their posttreatment observation period, and their footpads were harvested and plated for CFU. Footpad tissue was harvested after thorough disinfection with 70% alcohol swabs and then homogenized by fine mincing before suspension in sterile PBS. Tenfold serial dilutions and undiluted fractions of homogenate were plated in 0.5-ml aliquots on selective 7H11 agar supplemented with 10% oleic acid-albumin-dextrose-catalase (OADC) and incubated at 32°C for up to 12 weeks before CFU were enumerated.

Resistance testing.

The proportion of Q203-resistant CFU among all CFU in the wild-type M. ulcerans strain used for infection and in isolates obtained from mice after Q203 treatment was estimated by plating serial 10-fold dilutions of a culture or colony suspension in parallel on drug-free 7H11 agar plates and plates containing 0.3 ng/ml (4× MIC) and 3 ng/ml (40× MIC) of Q203. The plates were incubated at 32°C and CFU were counted after 10 weeks to determine the proportion of resistant CFU growing on Q203-containing plates. Mouse isolates were determined to be Q203-resistant if the proportion obtained by dividing the CFU count on Q203-containing plates by the CFU count on drug-free plates was ≥0.01.

For mutation analysis, DNA was extracted by boiling a few colonies in 100 μl of 1× TE (Tris 10 mM plus EDTA 1 mM) buffer. A 5-μl aliquot of this was used for amplification by PCR. The entire qcrB gene, along with 150-bp flanking regions, was amplified using specific primers (Table S6). The resultant 2,000-bp product was sequenced to identify the mutations in the gene.

Statistical analysis.

GraphPad Prism was used for statistical analysis. When comparing mean CFU counts between three or more groups within the same mouse strain, one-way analysis of variance was used with Dunnett’s or Bonferroni’s posttest to adjust for multiple comparisons. When data were not normally distributed, as with relapse time points at which some mice had 0 CFU and others had rebounding CFU counts, creating a bimodal distribution, group CFU counts were compared using the nonparametric Kruskal-Wallis test with Dunn’s test to adjust for multiple corrections. Similarly, an unpaired t test or Mann-Whitney test was used as the parametric or nonparametric test when comparing CFU counts between two groups. Two-way analysis of variance was used with Bonferroni’s posttest to test for interactions between treatment regimen and mouse strain.