ABSTRACT
Mycobacterium abscessus (Mab) is a rapidly-growing nontuberculous mycobacterium that is a growing health concern among both immunocompetent and immunocompromised patient populations. It most commonly causes skin and soft tissue or pulmonary infection. As an emerging health issue there is much that still needs to be understood about the infection and its progression to disease. In the context of pulmonary infection, an in vivo system of Mab infection that permits investigation of host-microbe interactions that result in Mab infection and the transition to pathogenesis, and also the evaluation of treatments, is an essential tool that is currently lacking. Here, we describe a system of pulmonary Mab infection in the C3HeB/FeJ mouse strain under corticosteroid immunosuppressive therapy that progresses to pathology.
INTRODUCTION
Mycobacterium abscessus (Mab) is a rapidly-growing nontuberculous mycobacterium. This environmental mycobacterium, while often an innocuous colonizer, can cause pathology in the lungs, skin, and other soft tissues. In patients with underlying immunodeficiency and/or lung conditions such as cystic fibrosis or bronchiectasis, Mab can cause a chronic lung infection1,2. Despite the emerging significance of Mab pulmonary infections, a mammalian model that recapitulates the process of aerosol infection and subsequent development of pathology in human disease3–5 has not yet been developed. Such a model would be an essential tool to study pathogenesis of Mab disease and to assess efficacy of treatment regimens.
In recent years, several attempts have been made to develop models of Mab infection. Models using hosts such as Drosophila6,7, zebrafish8–13, and Galleria mellonella larvae14 have been described, but these models do not capture key aspects of pulmonary infection. As mouse models have been used extensively for the study of pulmonary M. tuberculosis infection and disease15–18, it is relevant to assess if the mouse can serve as a host to develop a model of pulmonary Mab disease that is representative of aspects of human infection and pathology. We therefore turned to the mouse to develop a system of pulmonary Mab infection that utilizes the aerosol route for infection, and leads to an increase in lung bacterial burden after implantation resulting in the formation of gross lung lesions. Several studies have generated significant insights into infection dynamics of Mab in various genetic backgrounds of mice and demonstrated that immunocompetent mice gradually clear the infection, while certain immunodeficient mice can sustain chronic infection. The intravenous route of infection used in four of these studies19–22 represents a septicemia model and therefore is unlikely to mimic early stages of host-pathogen interactions that occur during infection via the aerosol route. Similarly, inoculation of Mab bolus intratracheally23–25 does not adequately mimic the process of aerosol infection. This approach is also highly invasive, requiring survival surgery for each mouse. Using intratracheal infection, Byrd et al. reported that Mab sustains chronic infection in the lungs of SCID mice but is cleared rapidly in immunocompetent BALB/c mice23, while Choi et al. observed similar clearance in immunocompetent C57BL/6 mice24. Caverly et al. demonstrated a relatively stable lung bacterial burden in C57BL/6 mice using a modified fibrin plug model of intratracheal infection, but clearance of the infection did occur in “cystic fibrosis” S489X Cftr mutation mice25
Several papers have attempted to bypass these more invasive modes of infection in favor of the aerosolization method, with varying degrees of success. Ordway et al. observed that while immunocompetent guinea pigs and mice gradually cleared Mab infection, in immunocompromised IFN-γ knockout mice Mab established infection in the lungs and spleen and proliferated in these organs26. De Groote et al. demonstrated in the GM-CSF knockout mouse a relatively stable lung bacterial burden with dissemination to the spleen as well, but the development of lung pathology is likely impacted by the lack of this important immunological differentiation factor27. N’Goma et al. also utilized the aerosol route of infection to compare the virulence of Mab cocultured with amoebae in Balb/c mice, although they did not demonstrate any increase in lung bacterial burden over time in this immunocompetent strain28. The same lack of increase in lung bacterial burden after aerosol infection was observed by Le Moigne et al. in the “cystic fibrosis” ΔF508 mouse29.
Based upon these previous studies, we hypothesized that transient pharmacological-induced immune suppression of immunocompetent mice may allow proliferation of Mab during acute stage infection followed by a sustained infection that leads to development of lesions. As mentioned, development of pathology is a valuable metric of infection for our system because of its usefulness in future investigations of Mab virulence and host pathogen interactions. We therefore chose C3HeB/FeJ mice as the host based on the increasing evidence that, among immunocompetent mice, this strain most closely mimics the pathology of disease in humans from infection of a related mycobacteria, M. tuberculosis (Mtb)15–17,30.
Although C3HeB/FeJ mice are not inherently immunocompromised, we hypothesized that we could achieve a sustained infection with associated lesion formation in C3HeB/FeJ mice with the right immunosuppressive regimen. This model also allows for exogenous modulation of immune suppression and, therefore, is nimble to experimental needs. This hypothesis of the use of corticosteroids for immune suppression was based on epidemiological studies describing corticosteroid use as a risk factor for development of invasive infections with rapidly-growing mycobacteria31–33. Of note in in vivo studies, a study evaluating efficacy of antibiotic treatment against M. ulcerans, another nontuberculous mycobacteria, in mice in the setting of concurrent dexamethasone use found that corticosteroid use prior to infection resulted in increased bacterial burdens and higher degrees of pathology compared to control34. Additionally, studies using the Cornell model of latent tuberculosis35 have assessed rates of reactivation of latent tuberculosis in the setting of corticosteroid use36. We applied these concepts by administering corticosteroids to the mice to increase their susceptibility to an initial infection. With the combination of immune suppression and use of C3HeB/FeJ mice we hoped to design a sustained Mab pulmonary infection model with corresponding pathology development.
METHODS
Ethics
Animal procedures used in the following studies were performed in adherence to the Johns Hopkins University Animal Care and Use Committee and to national guidelines.
Bacterial Strains and Growth Conditions
Mycobacterium abscessus strain ATCC 19977 was used in all studies. Mab was cultured in Middlebrook 7H9 broth (Becton&Dickinson 271310) supplemented with 10% ADS (Albumin, Dextrose, NaCl) and 0.05% Tween-8037 with constant shaking at 220 RPM in an orbital shaker at 37°C. Organ homogenates were cultured on Middlebrook 7H11 selective agar (Becton&Dickinson 283810) supplemented with 10% ADS, 50 µg/mL cycloheximide (Sigma-Aldrich C7698), 25 µg/mL polymyxin B (Sigma-Aldrich P1004), 50 µg/mL carbenicillin (Sigma-Aldrich C1389) and 20 ug/mL trimethoprim (Sigma-Aldrich T7883). Quantitative and colony forming unit (CFU) counts were determined as previously described17, but with enumeration of plates after 5 days of incubation at 37°C.
Mice, infection and corticosteroid administration
Female Nu/Nu (Nude) mice, 4-5 weeks old, were procured from Charles River Laboratories. Female C3HeB/FeJ mice, 4-5 weeks old, were procured from Jackson Laboratories. All infections were performed by aerosolizing Mab culture using a Glas-Col Inhalation Exposure System, as per the manufacturer’s instructions. The infection cycle comprised of 15 minute of pre-heat, 30 minutes of nebulization, 30 minutes of cloud decay and 15 minutes surface decontamination. Corticosteroid regimens and study design for Study 1 and Study 2 are outlined in Tables 1 and 2, respectively. Cortisone (Sigma-Aldrich C2755) and dexamethasone (Sigma-Aldrich D1756) were dissolved in sterile 1x phosphate buffered saline, pH 7.4 (Quality Biological, 114-058-101) and administered by sub-cutaneous injection, 200 µL bolus using 26 gauge syringe, as described previously34,36,38,39.
Study 1
We used Nude and C3HeB/FeJ mice. C3HeB/FeJ mice were divided into two groups. The first group did not receive any corticosteroid. The second group received cortisone daily (7 days/week, 40 mg/kg) beginning two days prior to infection. A cortisone dose of 40 mg/kg and daily dosing frequency were selected based on published usage in Mtb infection models36,38. The athymic Nude mouse group acted as a comparison to observations reported in the literature19,20. All mice were infected simultaneously by aerosolizing 10 mL of Mab suspension,A600nm= 1.0. Lungs were dissected at twenty-four hours post-infection to enumerate the CFU implanted during infection and then at two, four and six weeks post-infection to determine dynamics of infection. A schematic of this study is outlined in Table 1.
Study 2
C3HeB/FeJ mice were infected with Mab with either a ‘low’ or ‘high’ dose. For low dose, 10 mL of Mab suspension at A600nmof 0.1 was used. For high dose, 10 mL of Mab suspension at A600nmof 1.0 was used. In both infection groups there were three corticosteroid regimen groups: no corticosteroid, cortisone 40 mg/kg, or dexamethasone 10 mg/kg. We included dexamethasone with the aim to identify the optimal agent for inducing immune suppression. Cortisone was administered daily beginning two weeks prior to infection, and ending two weeks post-infection. Dexamethasone was administered on Mondays, Wednesdays, and Fridays, beginning two weeks prior to infection, and ending two weeks post-infection. A single dose of cortisone or dexamethasone was administered three days post treatment cessation to taper off the steroid exposure in the mice. Mice were sacrificed twenty-four hours post infection to determine bacterial burden at implantation and then two and four weeks post-infection. CFU determinations in the liver and spleen were undertaken to assess the extent of dissemination and growth in these organs. A schematic of this study is outlined in Table 2.
RESULTS
Study 1
The burden of Mab in the lungs of mice are shown in Figure 1A. On the day after infection, similar Mab CFU counts were found in the lungs of the three groups of mice: Nude, untreated C3HeB/FeJ and C3HeB/FeJ treated with cortisone for two days. In the lungs of untreated C3HeB/FeJ mice, Mab burden declined gradually initially (mean CFU of 3.4 log10 to 2.8 log10 from day zero to week two) and then more rapidly after two weeks (mean CFU of 2.8 log10 to 0.3 log10 from week two to week four). Similarly, Mab burden also declined in the lungs of Nude mice, but at a more gradual rate. The mean CFU from day zero was 3.6 log10, while at week twoit was within one standard deviation at 3.5 log10, but then declined to 2.7 log10 at week four and 1.1 log10 at week six. In C3HeB/FeJ mice receiving cortisone, the lung burden was relatively stable for the first two weeks (mean CFU of 3.6 log10 at day zero and 3.9 log10 at two weeks), followed by a rapid increase thereafter to 4.8 log10 at week four and 6.6 log10 at week six. At week six one mouse that received cortisone developed a pulmonary lesion (Figure 1B).
Study 2
We initiated a new study with the aim to reproduce and verify Mab proliferation and lesion formation observed in the lungs of C3HeB/FeJ mice in Study 1. Compared to Study 1, we administered immunosuppressive treatment for a longer period prior to infection with the intention to decrease the variability in CFU burden observed at two weeks, which was thought to be due to incomplete immunosuppression at infection and in the weeks following.
Mab burden in the lungs are shown in Figure 2. These infections resulted in an initial pulmonary implantation of 3.9±0.11 log10 and 4.4±0.10 log10 CFU for low and high Mab doses, respectively, averaged across all groups. Two weeks after infection, we observed a rapid increase in bacterial burden in the lungs of mice treated with cortisone or dexamethasone. For cortisone regimen mice with low dose and high dose infection, this meant an increase to 6.8 log10 and 6.7 log10 mean CFU, and for dexamethasone regimen mice with low dose and high dose infection an increase to 8.7 log10 and 7.9 log10 CFU. In the untreated mice at this same timepoint we observed a decrease to a mean CFU of 2.8 log10 in the low dose infection and 4.1 log10 in the high dose infection. We also observed bacterial dissemination to the liver and spleen in some of the mice (Tables 3, 4, and Supplement Table 1).
Shortly after discontinuing treatment, three mice infected with a high dose of Mab and treated with cortisone died. Similarly, in groups infected with high and low doses of Mab and treated with dexamethasone, three and nine mice died, respectively. There were no deaths in mice infected with low dose Mab and treated with cortisone.
At the week four time point, we observed a decrease in the average Mab burden in the lungs of mice in all groups. In untreated mice, lung bacterial burden decreased further to a mean CFU of 1.6 log10 CFU for low dose infection and to a mean CFU of 2.0 log10 for high dose infection over this two week period. For cortisone regimen mice with low dose and high dose infection mean CFU decreased to 3.3 log10 and 4.4 log10, and for dexamethasone regimen mice with low dose and high dose infection they decreased to 6.8 log10 and 6.6 log10 mean CFU. The bacterial burden in the liver and spleen, however, continued to increase or remained stable from week two (Tables 3 & 4).
Of note there was an increase in the variability of the lung bacterial burden at this timepoint compared to earlier timepoints, specifically in the dexamethasone regimen groups with both low and high infection doses. Data detailing the weights of all mice in this experiment along with their respective bacterial burdens at the sampled sites (Supplement Table 1) indicates that there appears to be a split in CFU burden that corresponds inversely with the weight of the animal, i.e. lower weight mice had higher bacterial burdens. In particular the dexamethasone regimen group at the week four timepoint mice with lower weights had higher lung CFU burdens and vice versa, resulting in the observed variability.
Observations of gross pathology of the lungs of mice at week 4 are shown in Figure 3. In the mice infected with low and high doses of Mab but not treated, the gross appearance of their lungs resembles the lungs of uninfected age-matched mice. In the cortisone treatment groups, similar levels of pathology, in terms of size, number and appearances of lesions in the lungs, were observed in both high and low dose infection groups. The lesions in the lungs of dexamethasone treated mice were larger in size and more numerous. The main factor that differentiated the levels of gross pathology observed in the lungs of mice in high or low infection dose groups was that the lungs of mice in the high dose infection group developed such a degree of inflammation and necrosis that they had fused to the thoracic wall, which made their intact extraction difficult. The levels of lung pathology observed qualitatively appear to correspond with the levels of bacterial burden in these groups over the course of the infection.
DISCUSSION
Our aim was to develop a system of pulmonary Mab infection in mice that utilizes the aerosol route of infection, leads to an increase in lung bacterial burden after implantation, and develops pulmonary lesions due to infection. The studies of Mab infection that have been performed in mice thus far have unfortunately failed to achieve these goals in an immunocompetent mouse strain. The study presented here describes a sustained Mab pulmonary infection via aerosolization in an immunocompetent mouse strain with corticosteroid-induced immune suppression.
Our first study affirmed that cortisone treatment in mice allows for a gradual increase in bacterial burden (Figure 1A). However, there was large variability in the bacterial burden at two weeks post-infection and a more uniformly higher bacterial burden was not observed until the week four timepoint. The variability in bacterial burden suggested that a longer duration of corticosteroid treatment prior to infection may be beneficial in increasing the initial bacterial burden in the lungs. Alternatively, this variability may also be explained by the mouse strain itself. C3HeB/FeJ mice are known to have variable infection outcomes when infected with M. tuberculosis16,17. To test this first hypothesis, we extended the corticosteroid regimen prior to infection in our second study. Here we observed the CFU in the lungs of mice treated with either cortisone or dexamethasone rapidly increased by the week two timepoint and did demonstrate decreased variability at this timepoint. We also observed dissemination to the liver and spleen in corticosteroid regimen mice, which did not occur to the same extent in untreated mice. As noted previously, several of the corticosteroid regimen mice subsequently died following cessation of steroids. However, these mice subjectively appeared to be sicker, of lower body weight, and seemed likely to die from the infection even prior to stopping treatment suggesting that the level of immune suppression was perhaps too high, especially in the dexamethasone-treated group.
Following cessation of corticosteroid administration, mice may have experienced some level of immune reconstitution, as evidenced by the decreased bacterial burden in the lungs observed across all groups at the week four timepoint. The rate of clearance observed after stopping corticosteroid administration appears to be gradual enough that experiments investigating virulence of Mab or treatment regimens against Mab would be able to be implemented with this system to examine changes in bacterial burden over time. Interestingly, despite the decrease in lung CFU, the bacterial burden in the liver and spleen remained stable or continued to increase. Dissemination of Mycobacterium chelonae, another rapidly growing mycobacteria, to skin and soft tissue has been associated with chronic use of low dose corticosteroid31. Additional studies to optimize the immunosuppressive regimen and initial infectious dose are currently underway in order to maximize the period of time the mice maintain a high bacterial burden, while minimizing the mortality.
Although further studies are needed to validate an aerosolized mouse model of a sustained Mab pulmonary infection, these preliminary infection studies are quite promising, and may be used as a guide for other investigators working on the development of mouse models of Mab infection and disease.
ACKNOWLEDGMENT
This work was supported by the Cystic Fibrosis Foundation grant. ESR was supported by the NIH T32 AI007291. The content is solely the responsibility of the authors and does not necessarily represent the official views of the Cystic Fibrosis Foundation or the National Institutes of Health.