The genetically engineered autoluminescent M. tuberculosis H37Rv (AlRv) strain, which harbors the luxCDABE gene cluster within its genome, was preserved at −80°C (Yang et al., 2015). AlRv displays a consistent emission of blue-green light independently, a characteristic attributed to the enzymatic activity encoded by the luxAB genes. luxCDE genes are responsible for the continuous production of the substrate aldehyde necessary for bioluminescence. Remarkably, AlRv demonstrates drug susceptibility and growth rate comparable to that of the wild-type M. tuberculosis strain. The traditional quantification method of CFUs was replaced by the assessment of relative light units (RLUs), maintaining an approximate ratio of 10:1.

The AlRv was subcultured in Middlebrook 7H9 broth (Difco, Detroit, MI, United States), supplemented with 10% oleic-acid-albumin-dextrose-catalase enrichment medium (BBL, Sparks, MD, United States), 0.2% glycerol, and 0.05% Tween80 at 37°C. The AlRv culture was grown until the optical density, measured at 600 nm, reached 0.6–0.8, with a corresponding RLUs level of 5 × 10⁶/mL. The optical density was measured using a V-1000 spectrophotometer (AOE, Shanghai, China), and the RLUs were quantified using the GLOMAX 20/20 luminometer (Promega, Madison, United States).

The inhalation delivery model for this experiment was constructed using three first-line anti-TB drugs with high solubility in water: rifampicin (RIF), isoniazid (INH), and ethambutol (EMB). All pharmaceutical drugs were purchased from Meilun (Dalian, China) with a minimum purity threshold of ≥95%. RIF and INH were solubilized in sterilized water, serving as the solvent, to attain stock concentrations of 2 mg/mL, which were subsequently diluted to 0.5 and 0.25 mg/mL, respectively. EMB was similarly dissolved into sterilized water at its original concentration of 10 mg/mL, then further diluted to 2.5 and 0.625 mg/mL. The prepared solutions of the drugs were refrigerated at 4°C until administration.

The female BALB/c mice (Gempharmatech, Foshan, China), aged six to eight weeks, underwent an acclimatization period of five to seven days before the initiation of experimental procedures. Subsequently, all mice were exposed to 10 mL broth culture of AlRv using an inhalation exposure system (Glas-Col, Terre Haute, IN) (Yu et al., 2020; Yu et al., 2021).

After infection, the mice were randomly divided into treatment and control groups, each containing six mice. Anesthesia was administered on the day treatment commenced, which corresponded to the third day post-infection in Experiment 1 and the day after infection in Experiment 2. Each mouse was placed individually into a container containing 0.5 mL of isoflurane for anesthesia induction. It took around 10 seconds for the loss of consciousness. Subsequently, the live RLUs of mice were measured by aligning their chests with the detection position of the GLOMAX 20/20 luminometer. After detection of the live RLUs, six mice were euthanized via cervical dislocation and positioned on a foam board with large head pins to secure them, ensuring their abdomens were oriented upwards. The mice were sterilized by spraying 75% ethanol and dissected to obtain the lungs. The organs were then placed into 2 mL sterilized phosphate-buffered saline, washed once with phosphate-buffered saline, and homogenized using glass tissue grinders. The lung suspensions were diluted using a 10-fold serial dilution method. Subsequently, the RLUs of 0.5 mL of undiluted and 10-fold diluted lung suspensions were measured using the GLOMAX 20/20 luminometer. Based on the detected RLUs and the ratio of CFUs to RLUs, the CFUs of the lung suspension were briefly speculated upon. The 0.5 mL properly diluted lung suspension was plated on 7H10 plates containing trimethoprim, actidione, carbenicillin, and polymyxin B to prevent contamination. All plates were then incubated in a 37°C thermostatic incubator for 4 weeks, following which the CFUs were determined.

Each cohort of six mice underwent daily administration of either 4 mL of solvent or drug solutions utilizing the Tow Systems Nose-Only Exposure Units (Tow-Int, Shanghai, China), as previously described (Fan et al., 2022; Yang et al., 2021; Zhao et al., 2023). The device of administration is illustrated in Supplementary Figure S1. Mice were subsequently secured in a holder and connected to the exposure cabinet. The drug or solvent solution was injected into the atomizer, which uses high-frequency oscillation technology through microporous screens to generate nebulized aerosols with diameters ranging from 1 to 5 μm for inhalation. The drug exposure duration for each group was approximately 20–25 min until 4 mL of the solutions had been administered.

The study consists of two main experiments, delineated as Experiment 1 and Experiment 2. The experimental protocols remained largely consistent, with the primary variations involving alterations in pharmacological dosages and the timing of medication administration subsequent to infection. The detailed designs for Experiments 1 and 2 were provided in Tables 1, 2, respectively. The live RLUs of the mice in the solvent group were continuously monitored from the day 11 for Experiment 1 (or day 13 for Experiment 2) post-infection. Measurements of RLUs for all groups were performed and the treatment stopped once the live RLUs of the solvent-treated mice exceeded 400. The duration of administration for Experiments 1 and 2 was 14 and 15 days, and the final live RLUs were recorded respectively (Tables 1, 2). Subsequently, all mice were euthanized the following day to quantify lung RLUs and CFUs using the above-mentioned method. The plates were then incubated at 37°C for 4 weeks, and then CFUs were counted.

The CFUs data revealed that RIF at a concentration (mg/mL) of 2, INH at 0.5, and EMB at 10 demonstrated significant in vivo anti-mycobacterial efficacy, as evidenced by the results of Experiments 1 and 2. These three concentrations were subsequently selected for further pharmacokinetic studies.

The mice were acclimated for 5 days prior to the commencement of the experiment. Subsequently, a volume of 4 mL of each drug was administered to six mice per group for each timepoint via inhalation for a precisely defined duration of 25 min. Blood samples were collected after picking eyeballs of mice (anaesthetized by inhalation of isoflurane) at various time points: 0 min (no treatment), 5 min, 15 min, 30 min, 45 min, 1 h, 2 h, 4 h, 8 h, 16 h, and 24 h after inhalable administration. A minimum of 300 µL of blood was collected from each mouse. Subsequently, the blood samples were transferred to anticoagulant-coated tubes and gently agitated to prevent haemolysis, ensuring exposure to the anticoagulant on the inner walls of the tubes. The blood samples were then centrifuged at 4°C and 3,000 rpm for 10 min. The resulting supernatant was transferred to new centrifuge tube and stored at −80°C for subsequent analysis. The mice that had undergone blood sampling were subsequently euthanised using cervical dislocation. Lung tissues were then excised, and residual blood and tissue fluids were removed with tissue paper. Distilled water was added to lung tissue in a ratio of 1.0 g vs. 1.0 mL. The lungs were subsequently homogenized to achieve a homogeneous mixture. A mixed standard acetonitrile solution was prepared, containing 50 ng/mL of each compound: verapamil, warfarin and methylsulfonylurea. A volume of 100 μL of the plasma sample or lung suspension was added to a tube, followed by the addition of 300 μL of the acetonitrile solution containing the corresponding internal standard compound. The mixture was vortexed and shaken for 30 s, then centrifuged at 12,000 rpm at 4°C for 10 min. The supernatant was transferred and filtered through a 0.22-μm filter into a new tube, then preserved at −80°C. All samples were submitted to liquid chromatography-mass spectrometry analysis (Q Exactive Focus, Thermo Fisher) for determination of the concentrations of the drugs.

Log-transformed data using GraphPad Prism version 8.3.0 was used to determine drug efficacy through a two-way analysis of variance, with p < 0.05 as statistical significance.

In Experiment 1, the autoluminescent M. tuberculosis burden in mice lungs was 3.60 ± 0.12 log₁₀CFU/lung assessed on the third day post-infection, preceding the commencement of inhalable treatments (BT in Figure 1). The live RLUs of all treatment groups were significantly lower than those of the solvent group (Figure 1, p < 0.0001). Additionally, these live RLUs were comparable to those of the before-treatment (BT) group (Figure 1, p > 0.05). Furthermore, the lung RLUs of all drug-treated groups were substantially lower than those of the solvent group (Figure 1, p < 0. 0001). Specifically, the lung RLUs of the RIF or EMB-treated groups were nearly identical to those of the BT group (Figure 1, p > 0.05), while the INH-treated group exhibited slightly lower lung RLUs than the BT group (Figure 1, p < 0.05). Moreover, the lung CFUs of all treatment groups were markedly lower than those of the solvent group (Figure 1, p < 0.0001). The RIF group had higher lung CFUs, whereas the INH group had lower lung CFUs compared to the BT group (Figure 1, p < 0.01 and p < 0.0001, respectively). The lung CFUs of the EMB-treated group were almost similar to those of the BT group (Figure 1, p > 0.05). The data from experiment 1 indicated that the autoluminescence-based inhalation delivery method can be utilized to establish a non-invasive model for testing drugs activities in vivo, using the same batch of live mice within 18 days after infection. Furthermore, the bactericidal effectiveness of drugs can be further evaluated by comparing lung CFUs of the different treatment groups rather than RLUs, as lung RLUs of the BT group were close to the background values.

Based on the results of Experiment 1, two additional dosages of each drug were added in Experiment 2. The bacterial burden in mice lungs measured on the day treatment initiated in Experiment 2 were 3.20 ± 0.11 log₁₀CFU/lung (BT in Figure 2C), which approximately equivalent to those recorded in Experiment 1 (BT in Figure 1). The live RLUs of the groups treated with RIF 2, INH 2, INH 0.5, EMB 10, EMB 2.5, and EMB 0.625 mg/mL were significantly lower compared to those of the solvent group (Figure 2A, p < 0.0001, p < 0.0001, p < 0.01, p < 0.001, p < 0.001, and p < 0.05, respectively). Conversely, the live RLUs of RIF 0.5, RIF 0.125 and INH 0.125 mg/mL groups did not show significant differences from those of the solvent group (Figure 2A, p > 0.05). These findings were also largely supported by both the lung RLUs and CFUs (Figures 2B,C), except for the lung CFUs of the RIF at 0.5 mg/mL group, which were slightly significantly lower than those of the solvent group (Figure 2C, p < 0.05). In summary, these findings suggest a negative correlation between live RLUs and administered dosages, indicating that live RLUs could serve as an initial criterion for assessing the therapeutic efficacy.

The lung RLUs of the RIF 2 mg/mL group demonstrated a significant reduction compared to those in the solvent group, with a log₁₀RLUs/lung difference exceeding 1.5 and even similar to that of the BT group (Figure 2B, p < 0.0001 and p > 0.05, respectively). Conversely, the lung RLUs in the RIF 0.5 and 0.125 mg/mL groups were comparable to those in the solvent group, indicating a lack of evident anti-M. tuberculosis activity at these dosages (Figure 2B, p > 0.05). The lung RLUs of the INH 2 and 0.5 mg/mL groups were significantly lower than those of the solvent group (Figure 2B, p < 0.0001), as all differences in log₁₀RLUs/lung exceeded 1.5 and were comparable to those of the BT group (Figure 2B, p > 0.05). The lung RLUs of the INH 0.125 mg/mL group were similar to those of the solvent group (Figure 2B, p > 0.05). In comparison, the lung RLUs of the EMB 10 and 2.5 mg/mL groups exhibited a pronounced decrease compared to those of the solvent group, with log₁₀RLUs/lung differences exceeding 1.5 and 1, respectively (Figure 2B, p < 0.0001). Meanwhile, the lung RLUs of the EMB 10 mg/mL group were equivalent to those of the BT group (Figure 2B, p > 0.05), whereas the lung RLUs in the EMB 0.625 mg/mL group showed less significant differences from the solvent group (Figure 2B, p < 0.05). In conclusion, the lung RLUs results are consistent with those obtained from the live RLUs, providing further support for the utility of detecting live RLUs in determining drug activity.

The lung CFUs in the RIF 2 mg/mL group exhibited a significant reduction compared to those in the solvent group with a difference in log₁₀CFUs/lung exceeding 1.5 (Figure 2C, p < 0.0001). The lung CFUs in the RIF 0.5 and 0.25 mg/mL groups were approximately comparable to those in the solvent group (Figure 2C, p < and close to 0.05 and p > 0.05, respectively). Notably, no colonies were observed on the plates of the INH 2 mg/mL group, indicating possibly eradication of AlRv from the mice lungs, considering that only a fraction of the lung suspension from each mouse was plated. The CFUs of the INH 0.5 mg/mL group were also significantly lower than those of the solvent group, with a difference of log₁₀CFUs/lung exceeding 1.5 and even lower than that of the BT group (Figure 2C, p < 0.0001). Moreover, all EMB dosages resulted in significantly reduced lung CFUs compared to the solvent control (Figure 2C, p < 0.0001, p < 0.0001, and p < 0.001, respectively). However, only the log₁₀CFUs/lung difference between the EMB 10 mg/mL and solvent group exceeded 1.5, with the lung CFUs in the EMB 10 group being nearly identical to those in the BT group (Figure 2C, p > 0.05). These CFU results were consistent with the findings derived from live or lung RLUs data. Overall, based on the comprehensive experimental data presented, we have successfully established an autoluminescence-based inhalation administration murine model for evaluating drugs activities against M. tuberculosis in vivo.

The concentrations of RIF, INH, and EMB in plasma and lung are illustrated in Figure 3. Pharmacokinetic properties of these three drugs in plasma and lung are summarized in Table 3.

The peak concentration of RIF was observed in both plasma and lung tissues 5 min post-administration following a single inhalation dose of 2 mg/mL, with values of 16.65 ± 3.91 μg/mL and 2.33 ± 0.89 μg/g, respectively (Table 3; Figure 3A). The maximum concentration of INH was observed in both plasma and lung samples, reaching values of 2.66 ± 0.29 μg/mL and 2.52 ± 0.81 μg/g, respectively, within 5 and 15 min post-inhalation at a dose of 0.5 mg/mL (Table 3; Figure 3B). However, INH was found to be undetectable in the lungs of mice 30 min post-administration (Figure 3B). The peak concentration of EMB was reached within 5 min following inhalation administration at a dose of 10 mg/mL, in both plasma and lung (Figure 3C). The peak concentrations of EMB were 28.64 ± 16.80 μg/mL in plasma and 10.94 ± 1.63 μg/g in lung (Table 3; Figure 3C).

The area under the concentration-time curve from zero to the maximum sampling time (AUC_0-t,), which was 4 h for INH and 24 h for RIF and EMB, was calculated to be 18.99, 1.57 and 103.15 μg h/mL in plasma, respectively (Table 3). The area under the concentration-time curve from zero to infinity (AUC_0-∞) for RIF, INH, and EMB in plasma was 20.97, 1.79, and 103.36 μg h/mL, respectively (Table 3). The mean residence time (MRT_0-∞) of RIF, INH, and EMB in plasma was 9.44, 0.85, and 5.11 h, respectively (Table 3). The half-life (t_1/2) of RIF, INH, and EMB in plasma was determined to be 6.27, 0.58, and 2.83 h, respectively (Table 3).

Thirty minutes post-administration, INH concentrations in the lungs of mice were found to be nearly undetectable, thereby precluding the determination of certain pharmacokinetic parameters, including AUC_0-∞, MRT_0-∞, and t_1/2. The AUC_0-t of INH in the lungs of mice was 0.34 μg h/mL (Table 3). The AUC_0-t, AUC_0-∞, MRT_0-∞, and t_1/2 of RIF in the lungs of mice were 17.47 ng·h/mL, 20.61 ng·h/mL, 12.76 h, and 8.23 h, respectively (Table 3). The AUC_0-t, AUC_0-∞, MRT_0-∞, and t_1/2 of EMB in the lungs of mice were 47.29 ng·h/mL, 48.70 ng·h/mL, 6.38 h, and 4.56 h, respectively (Table 3).