Interleukin-13-Overexpressing Mice Represent an Advanced Preclinical Model for Detecting the Distribution of Antimycobacterial Drugs within Centrally Necrotizing Granulomas

At late time points after aerosol infection with M. tuberculosis H37Rv, BALB/c IL-13^tg mice develop necrotic encapsulated granulomas (18), which are widely observed in human TB. Therefore, an optimal utilization of this mouse model for the correlation of lesion-specific pathology and drug penetration measured by MALDI-MS imaging requires sufficient time for necrotic pathology to develop. To assess the formation of granulomas after aerosol infection, lung tissue sections were histologically examined by 9 weeks postinfection. At this time point, an extensive pulmonary inflammation with cellular granulomas but also centrally necrotizing granulomas was observed (Fig. 1A). Cellular, inflammatory lesions were composed primarily of macrophages, lymphocytes, and neutrophils. For centrally necrotizing granulomas, a stringent stratification was evident, with an eosinophilic necrotic core consisting of dead and dying cells that was adjacent to regions of neutrophils (Fig. 1B and D), surrounded by a rim of foamy macrophages staining less intensely with eosin and positive for CD68 (Fig. 1D). A fibrous capsule demarcated the necrotic area from the adjoining tissue (Fig. 1C). Since the development of centrally necrotizing granulomas resembling the pathology of human TB was confirmed, drug treatment of BALB/c IL-13^tg mice started 9 weeks after aerosol infection.

To subsequently validate BALB/c IL-13^tg mice as an improved preclinical model for drug penetration studies into TB lesions, a workflow for sample preparation and analysis that considers safety regulations and provides drug distribution data had to be established (Fig. 2). To this end, M. tuberculosis-infected BALB/c IL-13^tg mice containing centrally necrotizing granulomas were treated with the drug combination of CFZ, PZA, and RIF, since the penetration of these drugs into human necrotic granulomas has been previously described by Prideaux et al. (17). Following therapy for a total of 10 consecutive days, mice were sacrificed and lung tissue was collected for later serial cryosections, whereby all work was conducted under biosafety level 3 (BSL-3) conditions. Analytical data are generally acquired in chemical laboratories (liquid chromatography-tandem MS [LC-MS/MS], MALDI-MS imaging) operating without biosafety level. Therefore, the inactivation of M. tuberculosis by gamma irradiation and its confirmation using the mycobacterial growth indicator tube (MGIT) system (26) were established. Only samples with associated culture-negative results were further processed. Also, the impact of the identified irradiation dose on tissue morphology, drug concentration, and drug distribution was investigated. The analysis of gamma-irradiated serial cryosections finally allowed a correlation of lung histopathology, drug distribution, and drug quantification.

Since previous studies applied gamma irradiation for the inactivation of mycobacteria prior to drug detection by LC-MS/MS or MALDI-MS imaging measurements (7, 13), this method was also used in the present study. In a first approach, frozen lung tissue biopsy specimens containing granulomatous lesions were exposed to different doses of gamma irradiation while frozen on dry ice, as radiation-induced damage of large molecules in hydrated materials is reduced at low temperatures (27). By applying a fixed position for the ¹³⁷Cs irradiation source and a rotation of the samples, an energy dose rate of 11.2 Gy/min was obtained. Subsequently, irradiated tissue was homogenized and inactivation of M. tuberculosis was evaluated by growth in liquid culture using the MGIT system, followed by detection of acid-fast bacilli (AFB) and cord formation in cultures. For homogenate generated from tissue that was not irradiated, a mean time to liquid culture positivity (TTP) of 4 days was determined (Table 1). Gamma irradiation of sample biopsy specimens with 16.10 or 32.20 kGy reduced the amount of viable M. tuberculosis, and the average time to positivity was 22 days or 19 days, respectively. However, even irradiation with 48.30 kGy did not completely eliminate M. tuberculosis, and in 2 out of 3 irradiated biopsy specimens, growth was detected, with a TTP value of 16 or 39 days. Since gamma irradiation of lung tissue biopsy specimens did not completely eradicate M. tuberculosis, in a second approach, cryosections were prepared and irradiated. To this end, lung tissue was cut into 12-μm-thick cryosections, which were mounted onto adhesive microscope slides and frozen at −80°C. During irradiation, sections were kept frozen on dry ice and a homogenous dose distribution was achieved by selecting an irradiation program with an oscillating irradiation source combined with a rotation of the specimens, resulting in an energy dose rate of 4.1 Gy/min. Subsequently, the sections were transferred from the microscope slides into liquid medium, and growth of M. tuberculosis was monitored using the MGIT system. For sections that were not irradiated, the mean TTP value was 14 days (Table 2). Irradiation of cryosections with 0.24 kGy, 0.49 kGy, or 0.98 kGy reduced the amount of viable M. tuberculosis so that a prolonged TTP was observed. After irradiation with 1.95 kGy, 3 out of 4 samples were MGIT negative and negative for AFB and only one sample was still TTP positive (18 days). Irradiation of cryosections with doses of 5.85 kGy, 11.71 kGy, or 17.56 kGy completely inactivated M. tuberculosis so that all MGIT tubes remained negative until day 84 of incubation and also no AFB were detected. Consequently, we identified that 5.85 kGy of gamma irradiation of cryosections is the lowest dose to inactivate M. tuberculosis H37Rv within centrally necrotizing granulomas of BALB/c IL-13^tg mice.

In order to investigate if gamma irradiation of cryosections has an effect on tissue morphology, serial cryosections from M. tuberculosis-infected BALB/c IL-13^tg mice were prepared, followed by irradiation of every other section with the dose of 5.85 kGy on dry ice, while the adjacent control sections were processed identically except for the irradiation procedure (see Fig. S1 in the supplemental material for a detailed description of sample preparation). Tissue integrity was assessed by histological and immunohistochemical staining (Fig. 3). Hematoxylin-eosin (HE) staining of irradiated or nonirradiated consecutive sections revealed no differences regarding the morphology of necrotic (Fig. 3A, n) and cellular regions of granulomas (Fig. 3A, c) or uninvolved tissue. Also, the fibrous capsule which demarcates the necrotic granuloma from the adjoining lung tissue was not altered after irradiation (Fig. 3B). Furthermore, immunohistochemical staining of CD68 showed that detection of CD68-expressing macrophages was similar in irradiated and nonirradiated adjacent cryosections, with macrophages located in cellular regions of granulomas as well as next to the fibrous rim of necrotizing granulomas (Fig. 3C), demonstrating that irradiation has no major impact on the integrity of cells such as macrophages.

In addition, the influence of gamma irradiation on the concentration of drugs within cryosections was investigated. For this purpose, naive BALB/c mice were treated with the CFZ/PZA/RIF combination for 5 consecutive days, and lung lobes were isolated and cut into 12-μm-thick serial cryosections. Frozen sections were irradiated (5.85 kGy) on dry ice, while the neighboring control sections were kept on dry ice without irradiation. In all sections, drug concentrations were quantified by LC-MS/MS analysis (see Fig. S1 for a detailed description of sample processing). The effect of gamma irradiation on the drug concentrations was determined by calculating the recovery rate for CFZ, PZA, and RIF in those sections that were irradiated in comparison to neighboring sections that were not irradiated (Fig. 4A). Concentrations of PZA were similar in cryosections irrespective of irradiation (Table S4), resulting in a mean recovery rate of 101% (standard deviation [SD], 23.17%), while the mean recovery rate of RIF within irradiated sections was slightly reduced to 94% (SD, 22.89%). Concentrations of CFZ were lower after irradiation with 5.85 kGy than in nonirradiated samples (Table S4), so that the mean recovery rate of CFZ was significantly reduced to 68% (SD, 17.63%) in irradiated cryosections. Since the concentration of CFZ was reduced after irradiation, we next investigated whether the distribution of CFZ is altered due to irradiation. To this end, drug distribution was analyzed by MALDI-MS imaging in adjacent lung cryosections that were either irradiated or not irradiated (see Fig. S1 for a detailed description of sample processing). Under both conditions, CFZ penetrated homogeneously throughout the healthy lung parenchyma and higher intensities were observed in close proximity to major airways that were distinguished by subsequent HE staining, demonstrating that irradiation has no impact on the distribution of CFZ (Fig. 4B).

Taking these findings together, we identified that an irradiation dose of 5.85 kGy eliminates M. tuberculosis within cryosections but affects neither tissue morphology nor the cellular composition of sections. Furthermore, gamma irradiation has no obvious impact on the concentrations of PZA and RIF. While the concentration of CFZ was reduced, the distribution of CFZ was not altered due to irradiation. Consequently, we successfully established a workflow for M. tuberculosis inactivation while preserving tissue morphology and staining properties, which are prerequisites for the detection of drug distribution by MALDI-MS imaging in conjunction with lesion-specific pathology.

By application of the established workflow, the impact of lesion heterogeneity on drug penetration was investigated in BALB/c IL-13^tg mice. To this end, therapy started 9 weeks after M. tuberculosis infection, mice were treated with the CFZ/PZA/RIF combination for 10 days, and 1 h after the last administration, lung tissue was collected and serial cryosections were prepared for histological characterization of lesions in combination with MALDI-MS imaging analysis. Since therapy started after necrotic pathology had developed, large granulomas consisting of necrotizing centers with karyorrhectic debris (Fig. 5A, n) that were surrounded by a collagen-rich fibrous layer (Fig. 5B) were readily apparent in addition to areas with cellular infiltration (Fig. 5A, c). A more detailed characterization of lesions by a combination of CD68 staining and Ziehl-Neelsen (ZN) staining revealed a rim of macrophages (Fig. 5C) on the inner side of the fibrotic capsule that were heavily infected with acid-fast rod-shaped M. tuberculosis (Fig. 5G). Besides these intracellular mycobacteria, high numbers of extracellular AFB were also observed within the necrotic center of granulomas (Fig. 5H to J), and here, even cording of mycobacteria, which is a characteristic feature of extracellular mycobacteria, was detected (Fig. 5H and I, red arrows).

MALDI-MS imaging analysis was performed in identical or neighboring cryosections to correlate the drug distribution with lesion pathology. The first-line drug PZA (Fig. 5D) exhibited a homogeneous distribution throughout the lung tissue and was detected within histopathologically normal lung tissue and areas with cellular infiltration as well as centrally necrotizing granulomas. The first-line drug RIF was present at high levels in uninvolved lung tissue and cellular areas, while lower signal intensities were detected within centrally necrotizing granulomas (Fig. 5E). The second-line drug CFZ was detectable within histopathologically normal lung tissue and in cellular inflammatory lesions, whereas only low intensities were measured within the core of large centrally necrotizing granulomas (Fig. 5F). Importantly, after the detection of PZA and CFZ (Fig. 5D and F) by MALDI-MS imaging, the same cryosection was reutilized for the histological characterization by CD68/ZN staining (Fig. 5C and G to J), allowing a direct comparison of drug distributions and target localizations. With this approach, we could clearly demonstrate that the first-line drug PZA penetrates tissue harboring intracellular and extracellular mycobacteria, while CFZ intensity is low in the center of large necrotizing granulomas, which, however, contains an enormous load of extracellular bacilli.

M. tuberculosis-infected BALB/c IL-13^tg mice develop heterogeneous pulmonary granulomas ranging from cellular inflammatory lesions through highly organized encapsulated granulomas with a cellular core to stratified centrally necrotizing granulomas (Fig. S2). Interestingly, depending on the lesion composition, CFZ abundance was higher in some cellular, inflammatory lesions (Fig. S2F), appeared to penetrate the rim of macrophages of smaller granulomas (Fig. S2F, stars), and was also detectable in the center of a stratified, encapsulated granuloma with a cellular core (Fig. S2A, encircled granuloma, and Fig. S2F). Regarding the distribution of the first-line drugs, PZA penetrated all lesion types (Fig. S2D), whereas the detection of RIF (Fig. S2E) was reduced in centrally necrotizing granulomas, confirming our observations described above.

While MALDI-MS imaging provides valuable information regarding the distribution of drugs within specific lesion compartments, absolute quantification of drugs is challenging. Therefore, drug concentrations within an entire neighboring cryosection were determined by LC-MS/MS measurements, yielding concentrations of 76 ng/mg for PZA, 25 ng/mg for CFZ, and 4 ng/mg for RIF.

As most preclinical efficacy studies for anti-TB drugs are performed in BALB/c mice, the pulmonary drug concentrations and distributions were also analyzed in BALB/c IL-13^tg-negative littermates by applying the same treatment regimen and sample preparation as that described for BALB/c IL-13^tg mice. Again, neighboring cryosections were used for histological characterization of lesions and determination of drug distributions and concentrations. The histological evaluation of pulmonary lesions revealed cellular, inflammatory lesions (Fig. 6A; Fig. S3A) composed predominantly of CD68-positive macrophages (Fig. 6C; Fig. S3C) with aggregates of lymphocytes interspersed throughout the lesion. These less-well organized granulomas were not separated by a collagen rim from the adjoining tissue (Fig. 6B; Fig. S3B) and did not show signs of necrosis. The drug distribution within these lesions was detected by MALDI-MS imaging in neighboring cryosections. The two first-line drugs PZA (Fig. 6D; Fig. S3D) and RIF (Fig. 6E; Fig. S3E) readily distributed in the cellular lesions of BALB/c mice and exhibited a relatively homogeneous penetration throughout the healthy tissue. Interestingly, CFZ was also detected throughout the section (Fig. 6F; Fig. S3F), but higher signal intensities were measured in those areas that correspond to cellular granulomas, which are composed primarily of large numbers of macrophages. Furthermore, drug concentrations were measured in neighboring cryosections by LC-MS/MS, resulting in concentrations of 37 ng/mg for PZA, 74 ng/mg for CFZ, and 6 ng/mg for RIF. Hence, with respect to the investigated drug distribution and concentration, the most striking differences between necrotizing granulomas and cellular granulomas were observed for CFZ.

IL-13^tg mice expressing murine il13 under the control of the human CD2 locus control region (72) [Tg(CD4-Il13)431Anjm] backcrossed to the BALB/c genetic background and littermates negative for the IL-13 transgene were bred under specific-pathogen-free conditions at the Christian-Albrechts-University of Kiel (Kiel, Germany) or the Max-Planck-Institute for Evolutionary Anthropology (Leipzig, Germany). M. tuberculosis-infected mice were maintained in individually ventilated cages (IVC; Ebeco, Castrop-Rauxel, Germany) under BSL-3 conditions at the Research Center Borstel. Female BALB/c mice (Charles River Laboratories) receiving only drug treatment were housed under standard conditions at the Research Center Borstel.

M. tuberculosis H37Rv was grown in Middlebrook 7H9 broth (Difco, Detroit, MI, USA) supplemented with 10% oleic acid-albumin-dextrose-catalase (OADC; Life Technologies, Gaithersburg, MD, USA), 0.05% Tween 80, and 0.2% glycerol (VWR International GmbH, Darmstadt, Germany). Mid-exponential-phase cultures were harvested, aliquoted, and frozen at −80°C. After thawing, the bacterial suspension was drawn through a nonpyrogenic needle (Microlance3; BD, Drogheda, Ireland) to ensure proper dispersion of mycobacteria prior to aerosol infection. PZA, CFZ, and RIF were purchased from Sigma-Aldrich (Taufkirchen, Germany). RIF was ground to a small particle size before sterile water was added (73). PZA was dissolved in sterile water and incubated at 55°C until particles were dissolved. CFZ was prepared by grinding with a mortar and pestle and added to 0.05% (wt/vol) agarose dissolved in sterile water (51). Concentrated stocks of individual drugs were kept at 4°C and combined on the day of application. PZA solutions, if particulate, were heated to 55°C before combination. Daily drug dosages of 150 mg/kg for PZA, 25 mg/kg for CFZ, and 10 mg/kg for RIF were chosen to achieve human-equivalent doses (44).

Mice were exposed to a low- or intermediate-dose aerosol infection with the M. tuberculosis H37Rv strain in a Glas-Col aerosol chamber, as previously described (74). Inoculum size was confirmed 24 h after infection by determining the bacterial burden in the lung. Treatment was initiated 9 weeks after infection, and the CFZ/PZA/RIF drug combination was administered by oral gavage for 10 consecutive days. At 1 h after the last administration, the mice were euthanized by CO₂ inhalation, lungs were aseptically removed, and lung tissue containing macroscopically visible lesions was snap-frozen in liquid nitrogen. BALB/c mice receiving only therapy without prior infection were treated with the CFZ/PZA/RIF combination by oral gavage for 5 consecutive days. Mice were sacrificed 0.5 h or 1 h after the last administration, and individual lung lobes were isolated and snap-frozen in liquid nitrogen. All animal experimentation was in accordance with the German Animal Protection Law and was approved by the animal research ethics committee of the federal state of Schleswig-Holstein (Germany) prior to permission by the Ministry of Energy, Agriculture, the Environment, Nature and Digitalization (Kiel, Germany; permit no. 3-1/15 and 69-6/16).

Serial cryosections (12 to 14 μm) were cut from unembedded lung tissue at −25°C using a Leica CM1850 or CM3050s cryostat (Leica Microsystems, Wetzlar, Germany) and thaw mounted onto adhesive glass slides (SuperFrost) as described by Treu et al. (12). Sections were stored at −80°C until further processing.

Further analysis of M. tuberculosis-containing samples required transfer to laboratories operating without biosafety level so that the inactivation of H37Rv by gamma irradiation had to be established. To this end, either lung tissue biopsy specimens or pulmonary cryosections frozen on dry ice were put into the irradiation chamber of a Biobeam 8000 (Gamma-Service Medical GmbH, Leipzig, Germany). The instrument is equipped with a ¹³⁷Cs gamma ray source (gamma energy = 0.662 MeV) with a half-life of 30.17 years and an activity of 81.21 TBq. Radiation doses were measured using alanine dosimetry. Elimination of M. tuberculosis was assessed by incubating the irradiated samples for at least 6 weeks in MGIT tubes (Becton, Dickinson GmbH, Heidelberg, Germany) at the National Reference Center for Mycobacteria at the Research Center Borstel. The MGIT contain 7 mL of modified Middlebrook 7H9 broth base for the cultivation of almost all mycobacteria, including M. tuberculosis. The complete medium contains OADC enrichment to optimize the growth of mycobacteria. Any positive sample was examined for AFB and cord formation.

To examine the development of centrally necrotizing granulomas, BALB/c IL-13^tg mice were sacrificed 9 weeks after infection, lung lobes were fixed in 4% formalin–phosphate-buffered saline (PBS), set in paraffin blocks, and sectioned (2 to 3 μm). Lung cryosections prepared from drug-treated BALB/c, BALB/c IL-13^tg, or BALB/c IL-13^tg-negative littermates were fixed in paraformaldehyde (PFA; 4% in PBS) (Morphisto, Frankfurt, Germany) or methanol. If MALDI-MS imaging analyses were performed prior to the immunohistochemical staining, the matrix was removed from the surface of the cryosection by gentle shaking in methanol. Standard protocols for HE and Heidenhain’s azan trichrome staining were used for histopathological analyses of paraffin-embedded sections or cryosections. For immunostaining, formalin-fixed and paraffin-embedded samples were deparaffinized, rehydrated, and subjected to a heat-induced epitope retrieval in the presence of 10 mM citric acid (pH 6.0). After cooling, slides were washed with TBS (50 mM Tris saline buffer, pH 7.6), and endogenous peroxidase activity was blocked by incubation with 3% H₂O₂ for 10 min. Following washing, sections were incubated in antibody diluent (Zytomed Systems GmbH, Berlin, Germany) containing a primary antibody specific for CD68 (ab125212; Abcam, Cambridge, UK) at room temperature with gentle agitation. After washing, a horseradish peroxidase (HRP)-conjugated polymer-based detection system (ZytoChem-Plus kit anti-rabbit; Zytomed Systems GmbH, Berlin, Germany) was used, followed by visualization with the chromogen 3-amino-9-ethylcarbazole (AEC; Permanent AEC kit; Zytomed Systems GmbH, Berlin, Germany), both according to the manufacturer’s instructions. Sections were counterstained with Gill’s hematoxylin, dehydrated, and coverslipped with Pertex (Medite, Burgdorf, Germany) mounting medium. To visualize M. tuberculosis and CD68 in the same lung section, fixed cryosections were stained for AFB using a ZN carbol-fuchsin solution (Merck, Darmstadt, Germany). After washing, endogenous peroxidases and biotin were blocked, followed by incubation with primary antibody against CD68 (ab125212 or ab53444, clone FA-11; Abcam, Cambridge, UK). For detection, biotinylated goat anti-rabbit antibody (Jackson Immunoresearch, Suffolk, UK) or biotinylated rabbit anti-rat IgG (Vector Laboratories, Burlingame, CA, USA), followed by the Vectastain Elite ABC peroxidase kit (Vector Laboratories, Burlingame, CA, USA), was used. Peroxidase was revealed by StayYellow/HRP (Abcam, Cambridge, UK) as the chromogen, and sections were counterstained with methylene blue (Merck, Darmstadt, Germany). Slides were imaged using a BX41 light microscope (Olympus, Hamburg, Germany) and NIS-Elements software (Nikon, Badhoevedorp, The Netherlands). For whole sections, images were stitched with Image Composite Editor (Microsoft, Redmond, WA, USA).

Lung cryosections were shipped frozen from the Research Center Borstel to the University of Bayreuth for MALDI-MS imaging analysis and stored at −80°C. Tissue sections were transferred into a desiccator for 10 min prior to matrix application by using a semiautomatic pneumatic sprayer system. Matrix solutions were prepared as follows: 2,5-dihydroxybenzoic acid (DHB), 20 mg/mL (Sigma-Aldrich, Taufkirchen, Germany), or 2,6-dihydroxyacetophenone (DHAP) (Sigma-Aldrich, Taufkirchen, Germany), 10 mg/mL, in methanol (MeOH)-H₂O (1:1, vol/vol). DHB was acidified with 0.1% (vol/vol) trifluoroacetic acid. MALDI-MS imaging measurements were performed using an AP-SMALDI10 source (TransMIT GmbH, Gießen, Germany) equipped with a 337-nm (λ) N₂ laser operating at a repetition rate of 60 Hz, coupled to a Q Exactive HF (Thermo Fisher Scientific GmbH, Bremen, Germany) orbital trapping mass spectrometer (75, 76). Measurements were carried out in positive and negative ion modes with one scanning event and 30 shots per pixel at a mass resolution of 240,000 at m/z 200 full width at half maximum. All measurements were performed with a fixed C-trap injection time of 500 ms. Matrix clusters of known sum formula (77) and the deprotonated molecule of phosphatidylserine (38:4) (C₄₄H₇₇NO₁₀P⁻) were used for internal mass calibration. Conversion of proprietary Thermo RAW files to the open data format imzML (78) was performed using the Java-based open access software jimzMLConverter version 1.3 (79). Data visualization was performed using MSiReader version 1.0 (80) with a bin width of ±2.5 ppm. Extracted ion images were interpolated using the linear interpolation setting 2. Mass-to-charge ratios of drug compounds are given as the calculated theoretical mass. Mass measurement accuracies are given as the root mean square error (RMSE) of the δm/z values in parts per million of each individual spectrum containing the targeted ion within a ±2.5 ppm window of the theoretical mass in Table S1 in the supplemental material. Compound identification was based on accurate mass. To confirm the identification of the drugs, lung sections of BALB/c IL-13^tg mice, which were not dosed with the drug combination (CFZ/PZA/RIF), were analyzed as negative control samples as we described elsewhere (82).

Prior to preparation for LC-MS/MS measurements, cryosections were photographed to determine the area of the sections using Fiji ImageJ software. Based on the known thickness of the cryosections and by assuming a density of 1 g/cm³, the weights of the cryosections were estimated. To measure the drug concentration, the sections were detached from the glass slides by suspending them in sterile water and subsequently dried using a SpeedVac. Afterwards, samples were reconstituted with 25 μL LC-MS-grade water, 200 μL acetonitrile (ACN) containing 12.5 ng/mL reserpine, and 25 μL 1% formic acid (FA). Samples were incubated at room temperature for 10 min under constant shaking and afterwards centrifuged for 10 min at 15,000 × g at room temperature. Approximately 200 μL of the resulting supernatant was transferred to a 1.5-mL Eppendorf tube and recentrifuged under the same conditions. Afterwards, 50 μL supernatant was transferred in a vial (three aliquots per sample), and the injection volume was 5 μL for the LC-MS/MS analysis.

Lung homogenate samples from untreated mice, prepared as previously described (81), were dried in a SpeedVac and subsequently reconstituted in 100 μL LC-MS-grade water, 800 μL ACN, and 100 μL 1% FA. The mixture was incubated at room temperature for 10 min under constant shaking, before the samples were centrifuged for 10 min at 15,000 × g. The resulting supernatant, approximately 700 μL, was collected in a separate tube and centrifuged under the same conditions to avoid floating particles. As a reference for the quantification, reserpine solution (12.5 ng/mL in ACN) was added to all samples. Calibration standards were prepared by diluting stock solutions in lung tissue homogenate from naive mice with individual standard curve ranges of 0.001 to 0.25 μg/mL for CFZ, 0.001 to 1 μg/mL for PZA, and 0.0025 to 0.1 μg/mL for RIF.

LC-MS/MS analysis was performed using a Waters Micromass Quattro Premier XE triple-quadrupole mass spectrometer (Waters Corporation, Milford, MA, USA) coupled to an 1100 series high-performance liquid chromatography (HPLC) system (Agilent Technologies, Santa Clara, CA, USA) using electrospray ionization (ESI). Samples were kept at 4°C in the autosampler during batch acquisitions.

For LC separation, a SeQuant ZIC-HILIC column (Merck Millipore SeQuant; 2.1-mm inner diameter by 150-mm length with a 5-μm particle size and a pore size of 200 Å) at a column temperature of 30°C was used. One percent FA (solvent A) and ACN (solvent B) were used as the mobile phase. The gradient started at 90% B at a flow rate of 0.5 mL/min. After 1 min of isocratic conditions, the percentage of ACN was decreased to 2% B until minute four. At minute four, the flow rate was increased to 0.8 mL/min. The gradient was kept isocratic at 2% B with a flow rate of 0.8 mL/min for 6 min until minute 10. Afterwards, the percentage of ACN was reincreased to 90% B until minute 15 and the flow rate was redecreased to 0.5 mL/min after minute 19. The conditions were maintained for 1 min, so that the total run time was 20 min (Table S2).

Multiple reaction monitoring (MRM) was used to quantify all analytes, and the corresponding transitions are listed in Table S3. Electrospray ionization in the positive ion mode was utilized with the following source parameters. The cone gas flow and desolvation gas flow were set to 100 L/h and 800 L/h, respectively. The extractor voltage was 3.0 V. The capillary voltage was set to 3.0 kV, the source temperature to 90°C, and the desolvation gas temperature to 450°C. MassLynx 4.1 (Waters Corporation, Milford, MA, USA) was used for operating the platform, and TargetLynx was used for data analysis. Quantitation was based on external calibration using untreated murine lung tissue in reference to reserpine.

If applicable, statistical analysis was performed using Prism 8 (GraphPad Software, San Diego, CA, USA). Quantifiable data are expressed as the means of individual determinations and standard deviations (SD). For determining the recovery rates after irradiation for PZA, CFZ, or RIF, nonirradiated and irradiated cryosections were compared using the Wilcoxon signed rank test.