The mouse model used in these studies was chosen based on a prior report that deletion of the gene IFT88 using the Cre/loxP system lead to de novo development of bronchiectasis in adult mice without evidence of antecedent infection. The model utilizes a conditional mouse mutant for the gene IFT88 that encodes for a protein playing an important role in development of normal cilia. In Cre recombinase positive mice homozygous for the IFT88 floxed allele, treatment with tamoxifen reportedly lead to progressive development of bronchiectasis starting at 3 weeks following gene deletion with an approximate 80% reduction in IFT88 protein expression after 2-3 weeks. In contrast, mice with the IFT88 floxed allele lacking Cre recombinase had wild type levels of IFT88 protein after tamoxifen administration (12). For our studies, deletion of IFT88 was accomplished by breeding IFT88 floxed mice (B6.129P2- Ift88^tm1Bky/J, Jackson Laboratory; Bar Harbor, ME) with mice possessing a tamoxifen-inducible Cre recombinase expressed from the actin promoter (B6.Cg-Tg(CAG-Cre/Esr1*)5Amc/J, Jackson Laboratory). Approximately 28 days after birth, mice were genotyped (Transnetyx; Cordova, TN) for the floxed Ift88 allele (forward primer 5’-GCGGCTGCAGAGATCCA-3’; reverse primer 5’-GGTATTGTTAGGAAGTAGTAAAACATAA-3’), the wildtype Ift88 allele (forward primer 5’-AGTAAGCGGCTGCAGAGATC-3’; reverse primer 5’-AATTCTGGCTCTGAACACAATCC-3’) and Cre (forward primer 5’-TTAATCCATATTGGCAGAACGAAAACG-3’; reverse primer 5’-CAGGCTAAGTGCCTTCTCTACA-3’). After selection of Cre recombinase positive/IFT88 floxed mice (experimental group) and Cre recombinase negative/IFT88 floxed mice (control group), both experimental and control mice received tamoxifen (Sigma-Aldrich; St Louis, MO) dissolved in corn oil (2 mg/100 µl) as daily intraperitoneal injections for 5 days. Since both IFT88 control and IFT88 KO mice received tamoxifen, differences observed comparing these groups are not attributable to tamoxifen per se. In addition the short duration of tamoxifen treatment (5 days) relative to the total duration of the experiments (approximately 3 months after tamoxifen administration) argues against an ongoing effect in our experiments. Depending on the experiment, Mabs infection of mice occurred 6-12 weeks after completion of the tamoxifen treatment. In addition to bronchiectasis, the other structural abnormality previously noted in these mice after deletion of IFT88 is the development of renal cysts 6 months after induction of the Cre/lox system (13), which is beyond the time frame of our experiments. It has also been found that loss of IFT88 in mice results in alteration in feeding behavior leading to weight gain – this is felt to be due to loss of cilia in the hypothalamus involved in regulating appetite (14). We observed this effect in the Cre recombinase positive IFT88 floxed mice after receiving tamoxifen. Observation of daily food intake by age-matched Cre recombinase negative IFT88 floxed male and female mice for a week indicated that these mice ate approximately 2.0 grams of food (Taklad 2920X; Envigo; Indianapolis, IN) per day. Thus, all mice received this amount of food throughout the course of experiments to maintain stable weight.

The experimental protocol used in these studies was approved by Institutional Animal Care and Use Committee of the University of New Mexico Health Sciences Center.

The Mabs 390R rough strain used in this study has been previously described (5, 15–19). It has been determined to be Mycobacterium abscessus subspecies abscessus (UT Health, The University of Texas Health Sciences Center at Tyler, Barbara A. Brown-Eliot).

Working stocks were prepared by culturing 390R in Middlebrook 7H9 medium (Sigma-Aldrich; St Louis, MO) supplemented with oleic acid-albumin-dextrose-catalase (OADC; Sigma-Aldrich) and 0.5% Tween 80 (Sigma-Aldrich) for 25 ± 5 h at 37°C and 100 rpm. The culture was centrifuged for 5 min at 15,000g and 4°C and the supernatant was discarded. The pellet was re-suspended in 10 mL of PBS with 0.5% Tween 80. Large clumps were allowed settle for 20 min, and then the supernatant was transferred into a new tube with 10 to 15 0.5 mm glass beads (BioSpec Products; Bartlesville, OK) and vortexed vigorously. Aliquots of the resulting single cell suspension was flash frozen in a dry ice-ethanol bath and stored at -80°C. The concentration of the working stocks was determined by plating serial dilutions on Middlebrook 7H11 Agarose (Sigma-Aldrich).

Bacteria for lung inoculation were prepared using several different methods. For intranasal infection with a single cell suspension, the inoculum was prepared by serial dilution from the frozen working stock with DPBS. 50 μL of the inoculum was placed on the nares of lightly anesthetized mice and allowed to be inhaled. To compare intratracheal (i.t.) infection with single cell suspension and bacterial aggregates, a fresh culture was prepared and processed using a procedure similar to that described to produce the working stock. The culture after 28.5 h incubation contained a mixture of individual bacteria and aggregates and was delivered using a 22-gauge intravenous (i.v.) catheter (Terumo; Tokyo, Japan). The single cell suspension after removing large clumps and homogenizing with glass beads was delivered using a MicroSprayer aerosolizer (Penn-Century, Inc; Wyndmoor, PA). For both methods of i.t. delivery, mice were lightly anesthetized with isoflurane and placed onto a dosing platform. A small animal laryngoscope (Penn-Century, Inc) was used to visualize the epiglottis and aid in proper placement of the catheter and the microsprayer, and 50 μl of the inoculum was pushed into the lungs. For infection using Mabs embedded in agarose bead, agarose beads were prepared following the method previously described for preparation of Pseudomonas aeruginosa-embedded agarose beads with slight modifications (20). Briefly, 390R was cultured in enriched Middlebrook 7H9 medium for 26 to 27 h at 37° C. A single cell suspension was prepared in DPBS with 0.5% Tween 80. 1 mL of the suspension was mixed with 10 mL of sterile 2% (w/v) Middlebrook 7H11 agarose (Sigma-Aldrich) maintained at 48°C. The mixture was slowly poured into a 250 mL Erlenmeyer flask containing 200 mL of sterile mineral oil (Sigma-Aldrich) heated to 48°C and stirred at 500 rpm with magnetic stir bar. After 5 min, ice was added around the flask of oil to quickly cool the agarose down and the mixture was stirred at a slower speed for another 20 minutes. The beads were washed 4 to 5 times with DPBS, each time followed by a 20 min centrifugation at 16,274g to remove remaining mineral oil. The agarose beads were delivered using a 22-guage i.v. catheter as previously described.

At indicated time points, mice were euthanized by intraperitoneal (i.p.) injection of 5.9 mg sodium pentobarbital (Vortech Pharmaceuticals; Dearborn, MI) in 100 μL. Lung tissues were removed aseptically and homogenized in 1 ml of DPBS using a BeadBeater (Biospec Products). The lung homogenates were serially diluted in DPBS, and then 10 μL of each dilution were spotted onto 7H11 agarose plates. After incubation at 37°C for 4 days, colonies were counted under a dissecting microscope.

The left lung was inflated using a lung inflation apparatus filled with 10% neutral buffer formalin. After 24 h, the trachea was tied off and the tissue was placed in 70% ethanol. The tissues were trimmed and processed by routine methods and stained with hematoxylin and eosin. In initial experiments, a veterinary pathologist examined the slides and scored them for the presence of cilia, inflammation and Clara cell hyperplasia. In a subsequent experiment, the veterinary pathologist examined H&E stained slides and scored them for the presence of granulomas, inflammation and histiocytes (nonlymphoid mononuclear cells). The HALO morphometry system (Indica Labs; Albuquerque, NM) was also used to quantify the extent of inflammation with the system trained to recognize areas of heavy inflammatory cell infiltration. In addition, the HALO morphometry system was used to quantify lung airway size by assessing airway perimeter and area.

Mice were injected i.p. with 150 U of heparin and euthanized 10 min later by i.p. injection of sodium pentobarbital. The lungs were then immediately removed with the trachea, bronchus, heart and vasculature, with or without lavage, and then perfused with 10 mL DPBS. The lung lobes were cut into small pieces and enzymatically digested with 30 µg/mL DNAse I (Sigma-Aldrich) and 0.7 mg/mL collagenase Type IV (Sigma-Aldrich) in RPMI 1640 medium for 1 hour at 37°C. The digested tissues were pushed through a 70 µm cell strainer (Corning; Tewksbury, MA) and a nylon wool column (G. Kisker, DE). Red blood cells were lysed in 1 mL RBC Lysis Buffer (ThermoFischer; Waltham, MA) at room temperature for 5 min, washed with HBSS, and resuspended in 1 mL of RPMI 1640 medium with 5% fetal calf serum. The cell suspension was layered over 3 mL of 30% Percoll in PBS (GE Healthcare, Cranbury, NJ) and centrifuged at 800g for 20 min to remove cell debris. Cells were analyzed without any antigen restimulation. Cells were incubated with monensin for 4 hours, fixed, surface stained with CD4-PerCP-Cy5.5 (clone RM4-5), permeabilized and stained using an antibody cocktail for IFNγ-FITC (clone XMG1.2), IL-17-PE (TC11-18H10.1), and FoxP3-Alex647 (clone MF23) or a cocktail of the respective isotype controls (antibodies, fixation, permeabilization solutions from BD Biosciences). Flow cytometry data was acquired on a C6 Plus Accuri (BD Biosciences) cytometer and analyzed with FlowJo (Treestar, Ashland OR). Singlet and lymphocyte gates were used with dead cells and debris excluded. Trypan blue staining was performed on all cell preparations analyzed by the flow cytometer. Our gating strategy is shown in the Supplemental Figure .

Mice were euthanized by i.p. injection of sodium pentobarbital. The lungs were then immediately removed and a cannula was inserted into the trachea. The lungs were lavaged three times with 0.70 mL cold DPBS, and the bronchoalveolar lavage fluid (BAL) was pooled, frozen in a dry ice-ethanol bath and stored for subsequent cytokine measurements. Cytokine and chemokines were detected and quantified using magnetic beads coated with analyte-specific capture antibodies (R&D Systems; Minneapolis, MN) and BioPlex 200 Instrument (Bio-Rad; Hercules, CA).

For comparisons unpaired, two tailed test was performed using GraphPad Prism version 9.3.1 for Windows, GraphPad Software, San Diego, California USA, www.graphpad.com. If less than 3 values were available from one or both groups to be compared, statistical analysis was not performed.

Mabs can persist in the lungs of immunocompromised mice, and recent studies provide evidence that with a high titer infecting lung inoculum, and embedding the infecting inoculum in substrate the barriers to infection in the lungs of mice with normal lung airways can be overcome resulting in varying degrees of persistent infection (10, 21). There are no prior reports of Mabs pulmonary infection in mice with structurally abnormal lung airways.

In the earliest reported study of Mabs lung infection in mice, which we carried out (5), SCID mice that received a low inoculum of Mabs 390R single cell suspension intranasally developed persistent lung infection for 28 days, the duration of the experiment. We used this as our initial approach in this study with Cre recombinase positive/IFT88 floxed mice (experimental group; IFT88 KO) and Cre recombinase negative/IFT88 floxed mice (control group; IFT88 Ctrl). Our first experiment with intranasal inoculation of Mabs 390R single cell suspension prepared from frozen working stock did not result in persistent lung infection. At 15 days after infection, we recovered roughly the same number of bacteria that was in the original inoculum in 70% of both control and IFT88 KO mice. However, the remaining animals at Day 15 had cleared the lung infection, and no bacteria was recovered from either group at 30 days (data not shown). As part of this experiment, a veterinary pathologist assessed lung histopathology in control and IFT88 KO mice. IFT88 KO mice demonstrated loss of cilia, Clara cell hyperplasia indicative of lung remodeling, and an increased composite inflammation score compared to control mice ( Figure 1 ). We repeated this experiment using a low intranasal inoculum of Mabs 390R resulting in total lung deposition of 2.3 x 10³ total lung CFU (data not shown). At day 20 after infection, we only recovered between 50 and 250 total lung CFU from both the IFT88 KO and control mice, and total lung CFU in all mice were below the limit of detection by day 36 after infection. Parallel groups of uninfected mice were included in this experiment for analysis of total lung CFU, BAL fluid cytokine analysis and total lung digests for analysis of T cell subpopulations. T cell subset analysis and cytokine levels from this experiment ( Figures 2A and 3A ) are included with data from a separate experiment described below.

We hypothesized that aggregates of bacteria more closely mimic natural infection in which virulent Mabs cording variants (390R in our study) replicate as cords and form clumps (5, 15, 22, 23) and therefore might be more likely to induce persistent infection. To test this hypothesis, we used an overnight culture, which contained a mixture of individual bacteria and aggregates of various sizes, to infect mice. Despite achieving a high lung deposition, total lung CFU dropped 2-4 logs by day 12 with no significant difference compared to single cell suspension which had been prepared at the same time ( Figure 4A ). We next assessed Mabs growth in control mice using an agarose bead inoculum which has been reported to produce persistent Pseudomonas aeruginosa lung infection, and more recently, Mabs lung infection in mice with normal lung airways (10, 24, 25). A pilot experiment with a low Mabs agarose bead inoculum (average lung depositon of 1.7 x 10³ CFU) in IFT88 Ctrl mice showed persistent lung infection out to 30 days, the duration of the experiment ( Figure 4B ). Thus, in our next experiment using this Mabs agarose bead inoculum, we compared growth of Mabs in the lungs of control mice to growth in the lungs of IFT88 KO mice carried out to 54 days (approximately 8 weeks) post-infection ( Figure 4C ). Similar to Figure 4B , Mabs lung CFU increased in control mice at day 15, but fell below the limit of detection by Day 54. In contrast, 4 of 6 IFT88 KO mice had significant numbers of CFU detected in the lungs at day 54 ( Figure 4C ). Importantly, the difference in Mabs lung CFU between control mice and IFT88 KO mice in this model provides the basis for exploration of how abnormal lung airways influence the growth of Mabs in the lung. A key study, that formed the basis for our Mabs agarose bead inoculum (25), compared mouse BAL lung cytokines and inflammatory cells from mice receiving lung inocula of P. aeruginosa alone, P. aeruginosa embedded in agarose beads, agarose beads alone, or no treatment. The investigators reported that mice receiving P. aeruginosa either alone or embedded in agarose beads showed elevated BAL cytokines and increased BAL inflammatory cells compared to mice receiving no treatment. Importantly, mice receiving agarose beads alone did not differ from mice receiving no treatment in terms of cytokines or inflammatory cells. Thus, agarose is an immunologically inert component of the inoculum compared to the bacterial organism.

We also analyzed T cell subset in the lungs and cytokines in the BAL fluid in this experiment, and compared the results ( Figures 2B and 3B ) against the results from our earlier experiment in which we did not establish persistent infection with an intranasal inoculum of single cell suspension ( Figures 2A and 3A ). T cell subset analysis showed a statistically significant decrease in the percentage of CD4⁺FoxP3⁺ T cells in the total lung lymphocyte population of IFT88 KO mice compared to control mice in both uninfected and infected mice at D36 (90 days after receiving tamoxifen)( Figure 2A ), and at D54 (107 days after receiving tamoxifen) using the agarose bead inoculum ( Figure 2B ). The significantly decreased percentage of T regulatory cells in the total lung lymphocyte population of IFT88 KO mice indicates a shift toward a proinflammatory lung milieu compared to IFT88 control mice. Furthermore, the decreased percentage of CD4⁺FoxP3⁺ T cells in the total lung lymphocyte population of IFT88 KO mice is not dependent on infection with Mabs and is apparent 3 months after deletion of the IFT88 gene (treatment with tamoxifen).

IFT88 KO mice infected using the agarose bead inoculum in which persistent infection was established ( Figure 3B ) showed much higher absolute levels of proinflammatory cytokines compared to mice inoculated with single cell suspension in which persistent infection was not established ( Figure 3A ). Levels were highest at day 15 with return toward baseline at day 54, although in some cases differences between IFT88 KO and IFT88 Ctrl mice persisted at this later time point ( Figure 3B ). These increased levels likely reflect the persistent Mabs infection established using the agarose bead inoculum. Overall, these results highlight the heightened proinflammatory response demonstrated in IFT88 KO mice.

Of the proinflammatory cytokines assayed, IL-22 was noteworthy in that it was significantly elevated in IFT88 KO mice relative to IFT88 control mice in both the uninfected and Mabs-infected groups ( Figures 3A, B ), suggesting an important role for this cytokine in our infection model. Also noteworthy is the significant elevation of IL-6 in the lungs of infected IFT88 KO mice relative to infected control mice independent of the establishment of persistent infection, supporting the inherent proinflammatory phenotype of IFT88 KO mice ( Figures 3A, B ). In addition to IL-22, IL-6 may thus play a central in the pathogenesis of persistent infection in IFT88 KO mice. IL-17 was not significantly elevated in IFT88 KO mice compared to IFT88 control mice in either uninfected or infected mice in the absence of persistent infection ( Figure 3A ). In contrast, it was significantly elevated at day 15 in infected IFT88 KO mice compared to IFT88 Ctrl mice where persistent infection was established using the agarose bead inoculum at day 15 ( Figure 3B ). T cell subset analysis of CD4⁺IL-17⁺ T cells at this time point shows an increased percentage of these cells at day 15 although this difference did not reach statistical significance ( Figure 2B ). This suggests that the elevated level of IL-17 at this time point represents a secondary, exaggerated proinflammatory response in IFT88 KO mice that may have a direct bearing on the outcome leading to persistent infection at day 54.

Cytokines were assayed in both experiments, and in some cases additional cytokines were added for the experiment using the agarose bead inoculum ( Figure 3B ). Where specific cytokines were assayed in both experiments, data may be reported for only one experiment since values were below the detectable range of standard curves generated in the Luminex assay. This was the true for cytokines TNFα, IFNγ, and IL-2. Elevated levels of the proinflammatory cytokine TNFα were observed in infected IFT88 KO mice relative to control mice ( Figure 3B ) in keeping with the proinflammatory phenotype of these mice. The lack of detectable IFNγ in IFT88 Ctrl mice at days 15 and 54, as well as the non-significant transient elevation of IFNγ seen in at day 15 in IFT88 KO mice ( Figure 3B ) suggests that in this model the Th1 immune response is not playing a role in pathogenesis. This is in keeping with our T cell subset analysis. Finally, IL-2 was significantly elevated in infected IFT88 Ctrl mice compared to IFT88 KO mice at day 15 ( Figure 3B ). This finding is consistent with the role that IL-2 is known to play in expanding the population of CD4⁺FoxP3⁺ T cells, and is reflected in our T cells subset analysis in this experiment at day 54.

In the experiment using the agarose bead inoculum, H&E stained lung sections examined by a veterinary pathologist were graded for the presence of granulomas, histiocytes in the alveoli, and peribronchiolar/perivascular inflammation. Consistent with Mabs persistence in IFT88 KO mice, and the findings of our T cell subset analysis and BAL cytokine determinations, there was a statistically significant increase in all three parameters in IFT88 KO mice compared to IFT88 Ctrl mice at D54 ( Figure 5 ). We also scanned slides using the HALO morphometry system and found increased inflammatory cells infiltrates in the lungs of IFT88 KO mice compared to IFT88 control mice ( Figure 6 ), although these differences were not statistically significant. There was considerable variation in the extent of inflammation both among different lung lobes in a single mouse and among different mice within the same treatment group likely due to the low infecting inoculum and low level of bacterial persistence in this study. This variability is likely the reason that the differences did not reach statistical significance.

As part of this assessment, slides were scanned for measurement of lung airway perimeter and area. Although both airway area and airway perimeter were increased in IFT88 KO mice compared to IFT88 Ctrl mice, these differences did not reach statistical significance ( Figure 7 ). When analyzing airway area, we were not able to replicate the method of measurement previously described for these mice (12). It was not clear how the investigators determined the diameter of arteries and airways that often have highly irregular shapes.