To compare the efficacy of L or S in monotherapy, Mtb-infected C3HeB/FeJ (n=7) and BALB/c (n=4–6) mice received L (administered 5/7 days per week orally at 100 mg/Kg) or S (administered 3/7 days per week on alternate days via intrapulmonary aerosol delivery at 100 mg/Kg and 50 mg/Kg, respectively) for 4 weeks. At the end of treatment, the animals were sacrificed, and the CFU in the lungs and spleen was enumerated (Figure 1A–D). Treatment of C3HeB/FeJ mice with L (7.30±0.45 log₁₀) or S (6.90±0.48 log₁₀) for 4 weeks resulted in an average of 0.51 and 0.91 log₁₀ CFU reduction in the lungs respectively, compared to untreated (UnRx) control (7.81±0.36 log₁₀) (Figure 1A). This difference failed to achieve statistical significance owing to the larger standard deviation associated with heterogenous advanced lung pathology observed in this model (Rathi et al., 2019; Irwin et al., 2015; Driver et al., 2012). In contrast, L (4.49±0.10 log₁₀) treated C3HeB/FeJ mice had significantly lower spleen bacterial burden compared to UnRx (5.43±0.27 log₁₀), with no significant difference between the L and S (4.94±0.11 log₁₀) treatment arm (Figure 1C). In BALB/c mice, L or S treatment was found to promote a significant reduction in lung bacterial burden (0.83 and 0.77 log₁₀, respectively) compared to UnRx or the vehicle-only controls (Figure 1B). There was no significant difference in lung bacterial burden after L or S treatment in BALB/c mice and there was no change in spleen bacterial burden compared to UnRx control (Figure 1D).

We further tested and compared the efficacy of L or S when used in combination therapy with BPa. Mtb-infected C3HeB/FeJ mice were treated with either BPaL (B=25 mpk; Pa = 100 mg/Kg and L=100 mg/Kg all administered 5/7 a week via oral gavage) or BPaS (BPa as in BPaL and S=100 mg/Kg administered 3/7 a week on alternate days via intrapulmonary aerosol delivery) for 4 weeks. The comparative analysis from combined data of three independent studies is shown in Figure 1E and G (data from individual studies are shown in Figure 1—figure supplements 1–3). Two of the three studies contained an extra group of BPa-treated mice as a reference control. Compared to UnRx (7.48±0.12 log₁₀) control, mice in the BPa (5.49±0.42 log₁₀), BPaL (4.76±0.38 log₁₀) and BPaS (4.98±0.35 log₁₀) treatment groups had significantly reduced the lung bacterial burden by 1.99, 2.72, and 2.50 log₁₀, respectively (Figure 1E). Although a higher CFU reduction was observed in the lungs of C3HeB/FeJ mice treated with BPaL or BPaS, these differences failed to achieve statistical significance compared to the BPa backbone regimen. All three regimens proved highly effective at reducing spleen bacterial burden in C3HeB/FeJ mice, with most mice returning no CFU within the limit of detection (LOD) employed herein (Figure 1G).

The effect of BPaL and BPaS (S=50 mg/Kg) combination therapy on the bacterial burden in the lungs and spleen of Mtb-infected BALB/c mice was determined at the end of 4 weeks of treatment. Figure 1F, H shows the combined lung and spleen CFU data from two independent studies (data for each study is shown in Figure 1—figure supplements 8–9). The combined result demonstrated that compared to UnRx (5.24±0.17 log₁₀) control, mice in the BPaL (1.29±0.13 log₁₀) and BPaS (1.44±0.17 log₁₀) treatment groups returned significantly fewer CFU in the lungs, with most mice returning no CFU within the LOD employed herein (Figure 1F). As in the C3HeB/FeJ TB model, no significant difference was observed in the lung CFU of BALB/c mice treated with either the BPaL or BPaS regimen. BPaL and BPaS therapy reduced BALB/c spleen bacterial burden to below the LOD of the assay with no CFU recovered for any treated mice (Figure 1H). In summary, these results support our hypothesis and demonstrate that both BPaL, BPaS (and BPa in the C3HeB/FeJ TB model) multidrug regimens show equivalent bactericidal effects in C3HeB/FeJ and BALB/c chronic TB efficacy models.

Five approaches were employed to monitor treatment-associated AEs in mice in this study including (1) changes in the body weight of mice; (2) lung histopathology and lesion scoring; (3) evaluation of complete blood count (CBC); (4) clinical pathology to study myelosuppression in the bone marrow and (5) changes in the content of immune cells in the lungs, spleen, bone marrow and blood.

No significant difference in the body weight among the treatment groups in either C3HeB/FeJ or BALB/c mice was observed following 4 weeks of treatment with S or L alone (Figure 1—figure supplements 4–7 and Figure 1—figure supplements 10–12, respectively). On the other hand, C3HeB/FeJ or BALB/c mice treated with BPa, BPaL, or BPaS showed marginal loss of body weight, ranging from 2.37–5.13% and only mice receiving the BPaL regimen, when compared to the UnRx control, reached statistically significant loses in body weight by the end of treatment (Figure 1—figure supplements 4 and 10, respectively).

The lesions in Mtb-infected UnRx C3HeB/FeJ mice (Figure 2A and B) showed a spectrum of diverse granuloma types ranging from aggregations of macrophages and lymphocytes to highly organized granulomas with collagen encapsulation and a region of central caseous necrosis that resembles to those found in some human patients (Irwin et al., 2015; Driver et al., 2012). By comparison, lung lesions of Mtb-infected UnRx BALB/c mice (Figure 2C and D) consisted of granulomas with a very homogeneous structure (Irwin et al., 2015) formed also by aggregations of macrophages and lymphocytes without a necrotic core. Mice treated for 4 weeks with the BPaL or BPaS regimen presented with a significant reduction in the number and size of granulomas in both C3HeB/FeJ and BALB/c TB models compared to their respective UnRx control, with no significant difference in lesion burden score between the combination drug treatment groups (Figure 2E and F).

The effect of L in the blood profile of humans and mouse has been reported (Tang et al., 2015; Gerson et al., 2002; Mase et al., 2022; Bigelow et al., 2020; Bigelow et al., 2021) but the same has not been reported for S. Therefore, a CBC profile was performed on Mtb-infected C3HeB/FeJ mice at the end of 4 weeks of treatment to quantify treatment-associated hematological changes. The results obtained from mice treated with L or S alone are summarized in Figure 3A. Of the 20-blood parameters evaluated, two blood parameters were affected during treatment. When compared to UnRx control and S-treated mice, L treatment significantly increased the red blood cell distribution width standard deviation (RDWs), while both L and S treatment were associated with a significant decrease in the mean corpuscular hemoglobin concentration (MCHC) compared to UnRx control (Figure 3A).

The Nix-TB trial associated the long-term administration of L within the BPaL regimen as the causative agent resulting in anemia in patients treated with the BPaL regimen (Conradie et al., 2020). Thus, the effect of combination therapy with the BPaL or BPaS regimen on CBC profile was analyzed at 1-, 2-, and 4 weeks of treatment (Figure 3B). None of the 20 parameters of CBC changed during the first 2 weeks of treatment. However, out of the 20 blood parameters evaluated, a total of four blood parameters were affected at 4 weeks of treatment. L-containing BPaL regimen was again associated with a significant increase in the RDWs, and lower hemoglobin (HGB) compared to UnRx control after 4 weeks. This effect was not observed in mice treated with BPa or BPaS (Figure 3B). However, as in monotherapy, there was a trend towards lower overall MCHC in mice treated with either BPaL or BPaS. The mean platelet volume (MPV) was marginally higher at 4 weeks in mice treated with BPa compared to BPaL or BPaS, albeit not significantly different from UnRx control (Figure 3B). Given that no difference in HGB or RDWs was observed between UnRx control and their comparator BPa and BPaS regimens, we concluded that the significant HGB decrease and RDWs increase (often observed during the development of anemia) were associated with inclusion of L in the BPaL regimen.

To further evaluate if L was associated with myelosuppressive effect, we performed hematopathology analysis on bone marrow from Mtb-infected C3HeB/FeJ and BALB/c mice at the end of treatment. For C3HeB/FeJ mice, the number of myeloid (M) and erythroid (E) precursor cells were calculated from H&E stained sections and their myeloid to erythroid ratio (M:E) was determined by counting 300 bone marrow cells in five different regions (Figure 3C). The BPa and BPaL treatment significantly decreased myeloid cells while increasing the proportion of erythroid cells (Figure 3D). Hence, the corresponding ratio in the bone marrow of animals treated with BPaL or BPa was lower compared to the UnRx control. Importantly, the BPaS treatment did not show any difference in the content of myeloid or erythroid cells when compared to UnRx control suggesting that S in the BPaS was able to recover this effect. In Mtb-infected BALB/c mice, the number of myeloid and other cell types were counted, however, no significant difference was found among the control and treatment groups (data not shown).

A comparative analysis for the concentration of cytokines and chemokines in the bone marrow, plasma, and lung samples from Mtb-infected C3HeB/FeJ mice treated with BPaL or BPaS regimen was also conducted. The bone marrow samples demonstrated a significant difference between the BPaL and BPaS groups, with appreciably higher level of pro-inflammatory cytokines and chemokines (IL-1β, IL-12p70, IL-23, TNFα, GROα (CXCL1), MP-2α (CXCL2), IP-10 (CXCL10), MP-1α (CCL3), RANTES (CCL5), MCP-3 (CCL7), and Eotaxin (CCL11)) in the BPaL- compared to BPaS-treated mice (Figure 4B). The plasma and lung samples, however, had similar cytokine and chemokine contents the treatment groups except for MCP-3 (CCL7) in plasma which was significantly higher in BPaS compared to the BPaL group (Figure 4—figure supplement 1). We also performed a correlation analysis of bone marrow cytokine and chemokine content with lung CFU obtained from treatment (BPaL or BPaS) and UnRx groups (Figure 4C). The analysis suggested that compared to UnRx control, there was a strong correlation between the profound reduction of lung bacterial burden and the profound reduction in bone marrow cytokine and chemokine contents observed in mice from BPaL or BPaS groups. Similar correlations were found between lung CFU and content of cytokines and chemokines in lung (Figure 4—figure supplement 2) and plasma (Figure 4—figure supplement 3).

We further assessed the environment of immune cells using flow cytometry in bone marrow, lungs, and blood from each group of Mtb-infected C3HeB/FeJ mice (Figure 5). The bone marrow (Figure 5A) results revealed that as compared to UnRx control, there was a significant decrease in the percentage of inflammatory myeloid phenotypes (CD45 +CD3-CD11b+CD11c-Ly6C+CCR2+, CD45 +CD3-CD11b+CD11c-Ly6C+CCR2+MHC-II+and CD45+CD3-D14+CCR2+) in response to therapy with either BPa, BPaL, or BPaS. In contrast, neutrophils (CD45 +CD3-CD11b+CD11c-Ly6C+Ly6G^high), precursor T cells (CD45 +CD3+), and B cells (CD45 +CD3 CD19+B220-) were significantly increased in either BPa, BPaL, or BPaS treatment groups compared to UnRx control in bone marrow. This reduced inflammatory response in treatment groups is also consistent in blood shown by significantly reduced inflammatory myeloid cells (CD45 +CD3-CD11b+CD11c-Ly6C+CCR2+) (Figure 5B). Interestingly, the response to therapy in the lungs (Figure 5C) was manifested by a significant increase of CD3 +CD4+T helper cells and B-1 cells (CD3-CD19+) and a reverse trend for CD3 +CD8+and γδ-T cells (CD3 +CD8+γδTCR+).

Furthermore, we also assessed changes in the distribution of immune cells in the lungs from Mtb-infected C3HeB/FeJ mice in response to therapy using multiplex fluorescent immunohistochemistry (mfIHC). A seven-color composite image for cell markers (B220, CD4, CD8, Foxp3, F4/80, and Ly6G) along with DAPI is shown in Figure 6A, while their single-color staining is shown in Figure 6B. Figure 6A shows a typical necrotic TB granuloma comprised of central necrosis, peripheral rim, and lung parenchyma. The analysis of mfIHC images revealed that the BPaL and BPaS treatments significantly and dramatically lowered the number of neutrophils (count based on Ly6G+) compared to UnRx control, however, F4/80+ cells were observed significantly higher in BPaS compared to UnRx control (Figure 6—figure supplement 1). Interestingly, the spearman’s correlation plot (Figure 6D) shows that a significant decrease in neutrophils was positively correlated with the corresponding increase in most of the other immune cells.

Bedaquiline fumarate (B) and linezolid (L) were obtained from LKT laboratories, pretomanid (Pa) from ChemShuttle, and 1599 (S: dihydrochloride) was provided by Dr. Lee at St. Jude Children’s Research Hospital. B was administered at 25 mg/kg. Pa and L were administered at 100 mg/kg each and S was delivered by inhalation in liquid form at 50 and 100 mg/kg to BALB/c and C3HeB/FeJ mice, respectively. The drugs were formulated in weekly batches according to the body weight of the animals, aliquoted for single daily dosing, and stored at 4 °C in the dark.

The drugs were prepared and administered as reported previously (Robertson et al., 2017; Gonzalez-Juarrero et al., 2021). Drug treatment was started 4 weeks post-aerosol infection for BALB/c and at 8–9 weeks post-aerosol infection for C3HeB/FeJ mice to allow time for lung pathology to fully develop. All drugs were administered once daily for 5 days/week for 4 weeks by oral (gavage) administration except S which was administered 3 days/week by intrapulmonary aerosol delivery using the Penn Century microsprayer as reported previously (Gonzalez-Juarrero et al., 2021). B was administered in the morning, Pa 1 hr after B and, L and S at least 4 hr after Pa.

After 4-weeks of treatment, C3HeB/FeJ and BALB/c mice were humanely euthanized by CO₂ narcosis. Blood, lungs, spleen, femur, and tibia bones were collected from each mouse for further processing and analysis.

The efficacy of the treatment was assessed by determining changes in bacterial burden [measured as CFU] in the lungs and spleen of animals at necropsy. The lungs and spleen were homogenized and prepared as reported previously (Gonzalez-Juarrero et al., 2021). The lung homogenates were plated onto 7H11 agar plates supplemented with 0.4% activated charcoal to reduce the carryover effect of drugs and incubated at 37 °C for 6–8 weeks before the final CFU count. The remaining lung homogenate was centrifuged, and the supernatant was collected and stored at –80 °C for evaluation of cytokines and chemokines.

The lungs were fixed in 4% paraformaldehyde (PFA) for 48 hr and then embedded in paraffin for histopathology. Sections from formalin-fixed and paraffin-embedded (FFPE) tissues were cut at 5 µm, stained with hematoxylin and eosin (H&E), and scanned at 40 X magnification using multispectral automated PhenoImager (Akoya Biosciences) for histopathological evaluation. The extent of lung lesion burden was quantified in blinded digital images using an open-source QuPath software for image analysis as described previously (Dutt et al., 2022). For each tissue section, a region of interest (ROI) was generated at low magnification with a custom tissue-detecting algorithm using decision forest training and classification to differentiate tissue versus background based on color and area. Lesions were identified within tissue ROIs at high magnification with an additional custom-made algorithm using decision forest training and classification based on staining intensity, color normalization and deconvolution, area, and morphological features. Percent lesion calculations were integrated into the same algorithm and calculated from tissue area and lesion area as designated by the ROI and lesions detected.

The sternum and one femur were fixed in 4% PFA and processed for histology. To evaluate the myelosuppressive effect of the drugs, bone sections were cut at 5 µm and stained with H&E. The number of myeloid and erythroid cells from 5 different regions of the bone were blinded and then counted by a veterinary clinical pathologist.

Briefly, a 0.6 mL sterile Eppendorf tube punctured at the bottom with the help of a 26-gauge needle was inserted into a 1.5 mL sterile Eppendorf tube. One end of the epiphysis of the long bones was cut open to expose the bone marrow and placed down into the small Eppendorf tube system. The tubes were centrifuged at 10,000 x g for 15 s and the marrow was collected from the base of the large Eppendorf tube. The bone marrow was resuspended in PBS and centrifuged again. Thereafter, the supernatant was collected and stored at –80 °C for evaluation of cytokines and chemokines while the bone marrow cells were saved in 4% PFA and freezing media for further use in clinical pathology analysis and flow cytometry, respectively.

For CBC analysis of C3HeB/FeJ animals during the treatment, blood was collected in K2-EDTA tubes via submandibular vein puncture as described previously (Golde et al., 2005). The blood was immediately analyzed in a VETSCAN HM5 hematology analyzer (Zoetis).

At the time of necropsy, whole blood was collected via cardiac puncture in K2-EDTA-containing tubes. After adding an equal volume of PBS, the samples were centrifuged at 800×g for 10 min at 25 °C with the brake off (deceleration = 0). The top plasma layer was collected and stored at –80°C for evaluation of their cytokine’s content. The buffy coat was collected, and washed and the erythrocytes were lysed using Miltenyi RBC lysis buffer (Miltenyi, CA). The cells were washed and resuspended in 500 µL of complete DMEM media and prepared for flow cytometry analysis.

Multiplex immunoassay was performed using a Luminex bead-based multiplex ELISA kit (ProcartaPlex Mouse Cytokine & Chemokine Panel 1 26plex, reference # EPXR260-26088-901, Invitrogen). Each sample was normalized to the total protein concentration determined by Bicinchoninic acid (BCA) assay (Thermo Fisher). The BCA and Luminex assay were performed according to the manufacturer’s instructions and the final stained samples were fixed with 4% PFA prior to acquisition. Sample data were acquired on a MAGPIX instrument running xPONENT 4.3 software (Luminex Corp.). Heatmaps were generated using the R pheatmap package. Correlation analysis of cytokine contents in bone marrow, plasma, and lungs with the lung bacterial burden was performed using the corrplot package in R.

Single-cell suspension of bone marrow, blood, and lungs from C3HeB/FeJ mice was prepared as described previously (Dutt et al., 2022). Cells counting, viability staining, and cell staining (Table 1) was performed accordingly (Dutt et al., 2022). Samples were acquired using Cytek Aurora 4-Laser spectral flow cytometer where 100,000 events were recorded. Data were analyzed in FlowJo software (BD Biosciences) using manual gating (Fox et al., 2020).

Five µm sections of FFPE lung tissues were stained for multiplex fluorescence immunohistochemistry (mfIHC) by the Imaging Core at the University of Colorado, Anschutz Medical Campus, Denver. The mfIHC was performed for a panel of 6-color antibodies + DAPI using the Opal-plex Tyramide Signal Amplification (TSA) technique using a Leica Bond III autostainer. The details of antibodies and Opal fluorophores used are given in Supplementary file 1. Each antibody was optimized using Opal 3-Plex Anti-Rb Detection Kit (Akoya Biosciences Inc cat# NEL830001KT) and stained with automated LabSat Research (Lunaphore Technologies SA, Epredia). Slides were scanned using multispectral automated PhenoImager (Akoya Biosciences) and analyzed for several immune cell populations using the inForm tissue Finder (Version 2.4.8) and Phenochart (Version 1.0.12) software (Akoya Biosciences).

Bacterial burden data were expressed as CFU which were Log₁₀-transformed and analyzed using GraphPad Prism version 9.5.1 (GraphPad Software, La Jolla, CA). The statistical analysis was performed using a Tukey’s multiple comparison test as part of either one-way or two-way ANOVA and mixed-model effect where necessary. The correlation analysis was performed using the spearman’s correlation test. Flow cytometry and mfIHC data were graphed in R studio and statistical evaluation was performed using stats package in R.